Org Mode Manual

This manual is for Org version 8.2.9.

Copyright © 2004–2017 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, with the Front-Cover Texts being “A GNU Manual,” and with the Back-Cover Texts as in (a) below. A copy of the license is included in the section entitled “GNU Free Documentation License.”

(a) The FSF's Back-Cover Text is: “You have the freedom to copy and modify this GNU manual.”

Introduction Getting started
Document Structure A tree works like your brain
Tables Pure magic for quick formatting
Hyperlinks Notes in context
TODO Items Every tree branch can be a TODO item
Tags Tagging headlines and matching sets of tags
Properties and Columns Storing information about an entry
Dates and Times Making items useful for planning
Capture - Refile - Archive The ins and outs for projects
Agenda Views Collecting information into views
Markup Prepare text for rich export
Exporting Sharing and publishing notes
Publishing Create a web site of linked Org files
Working With Source Code Export, evaluate, and tangle code blocks
Miscellaneous All the rest which did not fit elsewhere
Hacking How to hack your way around
MobileOrg Viewing and capture on a mobile device
History and Acknowledgments How Org came into being
GNU Free Documentation License The license for this documentation.
Main Index An index of Org's concepts and features
Key Index Key bindings and where they are described
Command and Function Index Command names and some internal functions
Variable Index Variables mentioned in the manual

Detailed Node Listing

Summary Brief summary of what Org does
Installation Installing Org
Activation How to activate Org for certain buffers
Feedback Bug reports, ideas, patches etc.
Conventions Typesetting conventions in the manual
Document structure
Outlines Org is based on Outline mode
Headlines How to typeset Org tree headlines
Visibility cycling Show and hide, much simplified
Motion Jumping to other headlines
Structure editing Changing sequence and level of headlines
Sparse trees Matches embedded in context
Plain lists Additional structure within an entry
Drawers Tucking stuff away
Blocks Folding blocks
Footnotes How footnotes are defined in Org's syntax
Orgstruct mode Structure editing outside Org
Org syntax Formal description of Org's syntax
Visibility cycling
Global and local cycling Cycling through various visibility states
Initial visibility Setting the initial visibility state
Catching invisible edits Preventing mistakes when editing invisible parts
Global and local cycling
Initial visibility Setting the initial visibility state
Catching invisible edits Preventing mistakes when editing invisible parts
Built-in table editor Simple tables
Column width and alignment Overrule the automatic settings
Column groups Grouping to trigger vertical lines
Orgtbl mode The table editor as minor mode
The spreadsheet The table editor has spreadsheet capabilities
Org-Plot Plotting from org tables
The spreadsheet
References How to refer to another field or range
Formula syntax for Calc Using Calc to compute stuff
Formula syntax for Lisp Writing formulas in Emacs Lisp
Durations and time values How to compute durations and time values
Field and range formulas Formula for specific (ranges of) fields
Column formulas Formulas valid for an entire column
Lookup functions Lookup functions for searching tables
Editing and debugging formulas Fixing formulas
Updating the table Recomputing all dependent fields
Advanced features Field and column names, parameters and automatic recalc
Link format How links in Org are formatted
Internal links Links to other places in the current file
External links URL-like links to the world
Handling links Creating, inserting and following
Using links outside Org Linking from my C source code?
Link abbreviations Shortcuts for writing complex links
Search options Linking to a specific location
Custom searches When the default search is not enough
Internal links
Radio targets Make targets trigger links in plain text
TODO items
TODO basics Marking and displaying TODO entries
TODO extensions Workflow and assignments
Progress logging Dates and notes for progress
Priorities Some things are more important than others
Breaking down tasks Splitting a task into manageable pieces
Checkboxes Tick-off lists
Extended use of TODO keywords
Workflow states From TODO to DONE in steps
TODO types I do this, Fred does the rest
Multiple sets in one file Mixing it all, and still finding your way
Fast access to TODO states Single letter selection of a state
Per-file keywords Different files, different requirements
Faces for TODO keywords Highlighting states
TODO dependencies When one task needs to wait for others
Progress logging
Closing items When was this entry marked DONE?
Tracking TODO state changes When did the status change?
Tracking your habits How consistent have you been?
Tag inheritance Tags use the tree structure of the outline
Setting tags How to assign tags to a headline
Tag groups Use one tag to search for several tags
Tag searches Searching for combinations of tags
Properties and columns
Property syntax How properties are spelled out
Special properties Access to other Org mode features
Property searches Matching property values
Property inheritance Passing values down the tree
Column view Tabular viewing and editing
Property API Properties for Lisp programmers
Column view
Defining columns The COLUMNS format property
Using column view How to create and use column view
Capturing column view A dynamic block for column view
Defining columns
Scope of column definitions Where defined, where valid?
Column attributes Appearance and content of a column
Dates and times
Timestamps Assigning a time to a tree entry
Creating timestamps Commands which insert timestamps
Deadlines and scheduling Planning your work
Clocking work time Tracking how long you spend on a task
Effort estimates Planning work effort in advance
Relative timer Notes with a running timer
Countdown timer Starting a countdown timer for a task
Creating timestamps
The date/time prompt How Org mode helps you entering date and time
Custom time format Making dates look different
Deadlines and scheduling
Inserting deadline/schedule Planning items
Repeated tasks Items that show up again and again
Clocking work time
Clocking commands Starting and stopping a clock
The clock table Detailed reports
Resolving idle time Resolving time when you've been idle
Capture - Refile - Archive
Capture Capturing new stuff
Attachments Add files to tasks
RSS Feeds Getting input from RSS feeds
Protocols External (e.g., Browser) access to Emacs and Org
Refile and copy Moving/copying a tree from one place to another
Archiving What to do with finished projects
Setting up capture Where notes will be stored
Using capture Commands to invoke and terminate capture
Capture templates Define the outline of different note types
Capture templates
Template elements What is needed for a complete template entry
Template expansion Filling in information about time and context
Templates in contexts Only show a template in a specific context
Moving subtrees Moving a tree to an archive file
Internal archiving Switch off a tree but keep it in the file
Agenda views
Agenda files Files being searched for agenda information
Agenda dispatcher Keyboard access to agenda views
Built-in agenda views What is available out of the box?
Presentation and sorting How agenda items are prepared for display
Agenda commands Remote editing of Org trees
Custom agenda views Defining special searches and views
Exporting Agenda Views Writing a view to a file
Agenda column view Using column view for collected entries
The built-in agenda views
Weekly/daily agenda The calendar page with current tasks
Global TODO list All unfinished action items
Matching tags and properties Structured information with fine-tuned search
Timeline Time-sorted view for single file
Search view Find entries by searching for text
Stuck projects Find projects you need to review
Presentation and sorting
Categories Not all tasks are equal
Time-of-day specifications How the agenda knows the time
Sorting agenda items The order of things
Filtering/limiting agenda items Dynamically narrow the agenda
Custom agenda views
Storing searches Type once, use often
Block agenda All the stuff you need in a single buffer
Setting Options Changing the rules
Markup for rich export
Structural markup elements The basic structure as seen by the exporter
Images and tables Images, tables and caption mechanism
Literal examples Source code examples with special formatting
Include files Include additional files into a document
Index entries Making an index
Macro replacement Use macros to create templates
Embedded LaTeX LaTeX can be freely used inside Org documents
Special blocks Containers targeted at export back-ends
Structural markup elements
Document title Where the title is taken from
Headings and sections The document structure as seen by the exporter
Table of contents The if and where of the table of contents
Lists Lists
Paragraphs Paragraphs
Footnote markup Footnotes
Emphasis and monospace Bold, italic, etc.
Horizontal rules Make a line
Comment lines What will *not* be exported
Embedded LaTeX
Special symbols Greek letters and other symbols
Subscripts and superscripts Simple syntax for raising/lowering text
LaTeX fragments Complex formulas made easy
Previewing LaTeX fragments What will this snippet look like?
CDLaTeX mode Speed up entering of formulas
The Export Dispatcher The main exporter interface
Export back-ends Built-in export formats
Export settings Generic export settings
ASCII/Latin-1/UTF-8 export Exporting to flat files with encoding
Beamer export Exporting as a Beamer presentation
HTML export Exporting to HTML
LaTeX and PDF export Exporting to LaTeX, and processing to PDF
Markdown export Exporting to Markdown
OpenDocument Text export Exporting to OpenDocument Text
Org export Exporting to Org
Texinfo export Exporting to Texinfo
iCalendar export Exporting to iCalendar
Other built-in back-ends Exporting to a man page
Export in foreign buffers Author tables and lists in Org syntax
Advanced configuration Fine-tuning the export output
HTML export
HTML Export commands How to invoke HTML export
HTML doctypes Org can export to various (X)HTML flavors
HTML preamble and postamble How to insert a preamble and a postamble
Quoting HTML tags Using direct HTML in Org mode
Links in HTML export How links will be interpreted and formatted
Tables in HTML export How to modify the formatting of tables
Images in HTML export How to insert figures into HTML output
Math formatting in HTML export Beautiful math also on the web
Text areas in HTML export An alternative way to show an example
CSS support Changing the appearance of the output
JavaScript support Info and Folding in a web browser
LaTeX and PDF export
LaTeX export commands How to export to LaTeX and PDF
Header and sectioning Setting up the export file structure
Quoting LaTeX code Incorporating literal LaTeX code
LaTeX specific attributes Controlling LaTeX output
OpenDocument Text export
Pre-requisites for ODT export What packages ODT exporter relies on
ODT export commands How to invoke ODT export
Extending ODT export How to produce ‘doc’, ‘pdf’ files
Applying custom styles How to apply custom styles to the output
Links in ODT export How links will be interpreted and formatted
Tables in ODT export How Tables are exported
Images in ODT export How to insert images
Math formatting in ODT export How LaTeX fragments are formatted
Labels and captions in ODT export How captions are rendered
Literal examples in ODT export How source and example blocks are formatted
Advanced topics in ODT export Read this if you are a power user
Math formatting in ODT export
Working with LaTeX math snippets How to embed LaTeX math fragments
Working with MathML or OpenDocument formula files How to embed equations in native format
Advanced topics in ODT export
Configuring a document converter How to register a document converter
Working with OpenDocument style files Explore the internals
Creating one-off styles How to produce custom highlighting etc
Customizing tables in ODT export How to define and use Table templates
Validating OpenDocument XML How to debug corrupt OpenDocument files
Texinfo export
Texinfo export commands How to invoke Texinfo export
Document preamble File header, title and copyright page
Headings and sectioning structure Building document structure
Indices Creating indices
Quoting Texinfo code Incorporating literal Texinfo code
Texinfo specific attributes Controlling Texinfo output
An example
Configuration Defining projects
Uploading files How to get files up on the server
Sample configuration Example projects
Triggering publication Publication commands
Project alist The central configuration variable
Sources and destinations From here to there
Selecting files What files are part of the project?
Publishing action Setting the function doing the publishing
Publishing options Tweaking HTML/LaTeX export
Publishing links Which links keep working after publishing?
Sitemap Generating a list of all pages
Generating an index An index that reaches across pages
Sample configuration
Simple example One-component publishing
Complex example A multi-component publishing example
Working with source code
Structure of code blocks Code block syntax described
Editing source code Language major-mode editing
Exporting code blocks Export contents and/or results
Extracting source code Create pure source code files
Evaluating code blocks Place results of evaluation in the Org mode buffer
Library of Babel Use and contribute to a library of useful code blocks
Languages List of supported code block languages
Header arguments Configure code block functionality
Results of evaluation How evaluation results are handled
Noweb reference syntax Literate programming in Org mode
Key bindings and useful functions Work quickly with code blocks
Batch execution Call functions from the command line
Header arguments
Using header arguments Different ways to set header arguments
Specific header arguments List of header arguments
Using header arguments
System-wide header arguments Set global default values
Language-specific header arguments Set default values by language
Header arguments in Org mode properties Set default values for a buffer or heading
Language-specific header arguments in Org mode properties Set language-specific default values for a buffer or heading
Code block specific header arguments The most common way to set values
Header arguments in function calls The most specific level
Specific header arguments
var Pass arguments to code blocks
results Specify the type of results and how they will be collected and handled
file Specify a path for file output
file-desc Specify a description for file results
dir Specify the default (possibly remote) directory for code block execution
exports Export code and/or results
tangle Toggle tangling and specify file name
mkdirp Toggle creation of parent directories of target files during tangling
comments Toggle insertion of comments in tangled code files
padline Control insertion of padding lines in tangled code files
no-expand Turn off variable assignment and noweb expansion during tangling
session Preserve the state of code evaluation
noweb Toggle expansion of noweb references
noweb-ref Specify block's noweb reference resolution target
noweb-sep String used to separate noweb references
cache Avoid re-evaluating unchanged code blocks
sep Delimiter for writing tabular results outside Org
hlines Handle horizontal lines in tables
colnames Handle column names in tables
rownames Handle row names in tables
shebang Make tangled files executable
tangle-mode Set permission of tangled files
eval Limit evaluation of specific code blocks
wrap Mark source block evaluation results
post Post processing of code block results
prologue Text to prepend to code block body
epilogue Text to append to code block body
Completion M-TAB knows what you need
Easy Templates Quick insertion of structural elements
Speed keys Electric commands at the beginning of a headline
Code evaluation security Org mode files evaluate inline code
Customization Adapting Org to your taste
In-buffer settings Overview of the #+KEYWORDS
The very busy C-c C-c key When in doubt, press C-c C-c
Clean view Getting rid of leading stars in the outline
TTY keys Using Org on a tty
Interaction Other Emacs packages
org-crypt Encrypting Org files
Interaction with other packages
Cooperation Packages Org cooperates with
Conflicts Packages that lead to conflicts
Hooks How to reach into Org's internals
Add-on packages Available extensions
Adding hyperlink types New custom link types
Adding export back-ends How to write new export back-ends
Context-sensitive commands How to add functionality to such commands
Tables in arbitrary syntax Orgtbl for LaTeX and other programs
Dynamic blocks Automatically filled blocks
Special agenda views Customized views
Speeding up your agendas Tips on how to speed up your agendas
Extracting agenda information Post-processing of agenda information
Using the property API Writing programs that use entry properties
Using the mapping API Mapping over all or selected entries
Tables and lists in arbitrary syntax
Radio tables Sending and receiving radio tables
A LaTeX example Step by step, almost a tutorial
Translator functions Copy and modify
Radio lists Sending and receiving lists
Setting up the staging area Where to interact with the mobile device
Pushing to MobileOrg Uploading Org files and agendas
Pulling from MobileOrg Integrating captured and flagged items

Next: , Previous: Top, Up: Top

1 Introduction

Next: , Previous: Introduction, Up: Introduction

1.1 Summary

Org is a mode for keeping notes, maintaining TODO lists, and doing project planning with a fast and effective plain-text system.

Org develops organizational tasks around NOTES files that contain lists or information about projects as plain text. Org is implemented on top of Outline mode, which makes it possible to keep the content of large files well structured. Visibility cycling and structure editing help to work with the tree. Tables are easily created with a built-in table editor. Org supports TODO items, deadlines, timestamps, and scheduling. It dynamically compiles entries into an agenda that utilizes and smoothly integrates much of the Emacs calendar and diary. Plain text URL-like links connect to websites, emails, Usenet messages, BBDB entries, and any files related to the projects. For printing and sharing notes, an Org file can be exported as a structured ASCII file, as HTML, or (TODO and agenda items only) as an iCalendar file. It can also serve as a publishing tool for a set of linked web pages.

As a project planning environment, Org works by adding metadata to outline nodes. Based on this data, specific entries can be extracted in queries and create dynamic agenda views.

Org mode contains the Org Babel environment which allows you to work with embedded source code blocks in a file, to facilitate code evaluation, documentation, and literate programming techniques.

Org's automatic, context-sensitive table editor with spreadsheet capabilities can be integrated into any major mode by activating the minor Orgtbl mode. Using a translation step, it can be used to maintain tables in arbitrary file types, for example in LaTeX. The structure editing and list creation capabilities can be used outside Org with the minor Orgstruct mode.

Org keeps simple things simple. When first fired up, it should feel like a straightforward, easy to use outliner. Complexity is not imposed, but a large amount of functionality is available when you need it. Org is a toolbox and can be used in different ways and for different ends, for example:

     • an outline extension with visibility cycling and structure editing
     • an ASCII system and table editor for taking structured notes
     • a TODO list editor
     • a full agenda and planner with deadlines and work scheduling
     • an environment in which to implement David Allen's GTD system
     • a simple hypertext system, with HTML and LaTeX export
     • a publishing tool to create a set of interlinked web pages
     • an environment for literate programming

There is a website for Org which provides links to the newest version of Org, as well as additional information, frequently asked questions (FAQ), links to tutorials, etc. This page is located at

The version 7.3 of this manual is available as a paperback book from Network Theory Ltd.

Next: , Previous: Summary, Up: Introduction

1.2 Installation

Org is part of recent distributions of GNU Emacs, so you normally don't need to install it. If, for one reason or another, you want to install Org on top of this pre-packaged version, there are three ways to do it:

We strongly recommend to stick to a single installation method.

Using Emacs packaging system

Recent Emacs distributions include a packaging system which lets you install Elisp libraries. You can install Org with M-x package-install RET org.

Important: you need to do this in a session where no .org file has been visited, i.e., where no Org built-in function have been loaded. Otherwise autoload Org functions will mess up the installation.

Then, to make sure your Org configuration is taken into account, initialize the package system with (package-initialize) in your .emacs before setting any Org option. If you want to use Org's package repository, check out the Org ELPA page.

Downloading Org as an archive

You can download Org latest release from Org's website. In this case, make sure you set the load-path correctly in your .emacs:

     (add-to-list 'load-path "~/path/to/orgdir/lisp")

The downloaded archive contains contributed libraries that are not included in Emacs. If you want to use them, add the contrib directory to your load-path:

     (add-to-list 'load-path "~/path/to/orgdir/contrib/lisp" t)

Optionally, you can compile the files and/or install them in your system. Run make help to list compilation and installation options.

Using Org's git repository

You can clone Org's repository and install Org like this:

     $ cd ~/src/
     $ git clone git://
     $ make autoloads

Note that in this case, make autoloads is mandatory: it defines Org's version in org-version.el and Org's autoloads in org-loaddefs.el.

Remember to add the correct load-path as described in the method above.

You can also compile with make, generate the documentation with make doc, create a local configuration with make config and install Org with make install. Please run make help to get the list of compilation/installation options.

For more detailed explanations on Org's build system, please check the Org Build System page on Worg.

Next: , Previous: Installation, Up: Introduction

1.3 Activation

Since Emacs 22.2, files with the .org extension use Org mode by default. If you are using an earlier version of Emacs, add this line to your .emacs file:

     (add-to-list 'auto-mode-alist '("\\.org\\'" . org-mode))

Org mode buffers need font-lock to be turned on: this is the default in Emacs1.

There are compatibility issues between Org mode and some other Elisp packages, please take the time to check the list (see Conflicts).

The four Org commands org-store-link, org-capture, org-agenda, and org-iswitchb should be accessible through global keys (i.e., anywhere in Emacs, not just in Org buffers). Here are suggested bindings for these keys, please modify the keys to your own liking.

     (global-set-key "\C-cl" 'org-store-link)
     (global-set-key "\C-cc" 'org-capture)
     (global-set-key "\C-ca" 'org-agenda)
     (global-set-key "\C-cb" 'org-iswitchb)

With this setup, all files with extension ‘.org’ will be put into Org mode. As an alternative, make the first line of a file look like this:

     MY PROJECTS    -*- mode: org; -*-

which will select Org mode for this buffer no matter what the file's name is. See also the variable org-insert-mode-line-in-empty-file.

Many commands in Org work on the region if the region is active. To make use of this, you need to have transient-mark-mode (zmacs-regions in XEmacs) turned on. In Emacs 23 this is the default, in Emacs 22 you need to do this yourself with

     (transient-mark-mode 1)

If you do not like transient-mark-mode, you can create an active region by using the mouse to select a region, or pressing C-<SPC> twice before moving the cursor.

Next: , Previous: Activation, Up: Introduction

1.4 Feedback

If you find problems with Org, or if you have questions, remarks, or ideas about it, please mail to the Org mailing list If you are not a member of the mailing list, your mail will be passed to the list after a moderator has approved it2.

For bug reports, please first try to reproduce the bug with the latest version of Org available—if you are running an outdated version, it is quite possible that the bug has been fixed already. If the bug persists, prepare a report and provide as much information as possible, including the version information of Emacs (M-x emacs-version <RET>) and Org (M-x org-version RET), as well as the Org related setup in .emacs. The easiest way to do this is to use the command

     M-x org-submit-bug-report RET

which will put all this information into an Emacs mail buffer so that you only need to add your description. If you re not sending the Email from within Emacs, please copy and paste the content into your Email program.

Sometimes you might face a problem due to an error in your Emacs or Org mode setup. Before reporting a bug, it is very helpful to start Emacs with minimal customizations and reproduce the problem. Doing so often helps you determine if the problem is with your customization or with Org mode itself. You can start a typical minimal session with a command like the example below.

     $ emacs -Q -l /path/to/minimal-org.el

However if you are using Org mode as distributed with Emacs, a minimal setup is not necessary. In that case it is sufficient to start Emacs as emacs -Q. The minimal-org.el setup file can have contents as shown below.

     ;;; Minimal setup to load latest 'org-mode'
     ;; activate debugging
     (setq debug-on-error t
           debug-on-signal nil
           debug-on-quit nil)
     ;; add latest org-mode to load path
     (add-to-list 'load-path (expand-file-name "/path/to/org-mode/lisp"))
     (add-to-list 'load-path (expand-file-name "/path/to/org-mode/contrib/lisp" t))

If an error occurs, a backtrace can be very useful (see below on how to create one). Often a small example file helps, along with clear information about:

  1. What exactly did you do?
  2. What did you expect to happen?
  3. What happened instead?
Thank you for helping to improve this program.
How to create a useful backtrace

If working with Org produces an error with a message you don't understand, you may have hit a bug. The best way to report this is by providing, in addition to what was mentioned above, a backtrace. This is information from the built-in debugger about where and how the error occurred. Here is how to produce a useful backtrace:

  1. Reload uncompiled versions of all Org mode Lisp files. The backtrace contains much more information if it is produced with uncompiled code. To do this, use
              C-u M-x org-reload RET

    or select Org -> Refresh/Reload -> Reload Org uncompiled from the menu.

  2. Go to the Options menu and select Enter Debugger on Error (XEmacs has this option in the Troubleshooting sub-menu).
  3. Do whatever you have to do to hit the error. Don't forget to document the steps you take.
  4. When you hit the error, a *Backtrace* buffer will appear on the screen. Save this buffer to a file (for example using C-x C-w) and attach it to your bug report.

Previous: Feedback, Up: Introduction

1.5 Typesetting conventions used in this manual

TODO keywords, tags, properties, etc.

Org mainly uses three types of keywords: TODO keywords, tags and property names. In this manual we use the following conventions:

TODO keywords are written with all capitals, even if they are user-defined.
User-defined tags are written in lowercase; built-in tags with special meaning are written with all capitals.
User-defined properties are capitalized; built-in properties with special meaning are written with all capitals.

Moreover, Org uses option keywords (like #+TITLE to set the title) and environment keywords (like #+BEGIN_HTML to start a HTML environment). They are written in uppercase in the manual to enhance its readability, but you can use lowercase in your Org files3.

Keybindings and commands

The manual suggests two global keybindings: C-c a for org-agenda and C-c c for org-capture. These are only suggestions, but the rest of the manual assumes that you are using these keybindings.

Also, the manual lists both the keys and the corresponding commands for accessing a functionality. Org mode often uses the same key for different functions, depending on context. The command that is bound to such keys has a generic name, like org-metaright. In the manual we will, wherever possible, give the function that is internally called by the generic command. For example, in the chapter on document structure, M-<right> will be listed to call org-do-demote, while in the chapter on tables, it will be listed to call org-table-move-column-right. If you prefer, you can compile the manual without the command names by unsetting the flag cmdnames in org.texi.

Next: , Previous: Introduction, Up: Top

2 Document structure

Org is based on Outline mode and provides flexible commands to edit the structure of the document.

Next: , Previous: Document Structure, Up: Document Structure

2.1 Outlines

Org is implemented on top of Outline mode. Outlines allow a document to be organized in a hierarchical structure, which (at least for me) is the best representation of notes and thoughts. An overview of this structure is achieved by folding (hiding) large parts of the document to show only the general document structure and the parts currently being worked on. Org greatly simplifies the use of outlines by compressing the entire show/hide functionality into a single command, org-cycle, which is bound to the <TAB> key.

Next: , Previous: Outlines, Up: Document Structure

2.2 Headlines

Headlines define the structure of an outline tree. The headlines in Org start with one or more stars, on the left margin4 5. For example:

     * Top level headline
     ** Second level
     *** 3rd level
         some text
     *** 3rd level
         more text
     * Another top level headline

Some people find the many stars too noisy and would prefer an outline that has whitespace followed by a single star as headline starters. Clean view, describes a setup to realize this.

An empty line after the end of a subtree is considered part of it and will be hidden when the subtree is folded. However, if you leave at least two empty lines, one empty line will remain visible after folding the subtree, in order to structure the collapsed view. See the variable org-cycle-separator-lines to modify this behavior.

Next: , Previous: Headlines, Up: Document Structure

2.3 Visibility cycling

Next: , Previous: Visibility cycling, Up: Visibility cycling

2.3.1 Global and local cycling

Outlines make it possible to hide parts of the text in the buffer. Org uses just two commands, bound to <TAB> and S-<TAB> to change the visibility in the buffer.

<TAB>     (org-cycle)
Subtree cycling: Rotate current subtree among the states
          ,-> FOLDED -> CHILDREN -> SUBTREE --.

The cursor must be on a headline for this to work6. When the cursor is at the beginning of the buffer and the first line is not a headline, then <TAB> actually runs global cycling (see below)7. Also when called with a prefix argument (C-u <TAB>), global cycling is invoked.

S-<TAB>     (org-global-cycle)
C-u <TAB>
Global cycling: Rotate the entire buffer among the states
          ,-> OVERVIEW -> CONTENTS -> SHOW ALL --.

When S-<TAB> is called with a numeric prefix argument N, the CONTENTS view up to headlines of level N will be shown. Note that inside tables, S-<TAB> jumps to the previous field.

C-u C-u <TAB>     (org-set-startup-visibility)
Switch back to the startup visibility of the buffer (see Initial visibility).
C-u C-u C-u <TAB>     (show-all)
Show all, including drawers.
C-c C-r     (org-reveal)
Reveal context around point, showing the current entry, the following heading and the hierarchy above. Useful for working near a location that has been exposed by a sparse tree command (see Sparse trees) or an agenda command (see Agenda commands). With a prefix argument show, on each level, all sibling headings. With a double prefix argument, also show the entire subtree of the parent.
C-c C-k     (show-branches)
Expose all the headings of the subtree, CONTENT view for just one subtree.
C-c <TAB>     (show-children)
Expose all direct children of the subtree. With a numeric prefix argument N, expose all children down to level N.
C-c C-x b     (org-tree-to-indirect-buffer)
Show the current subtree in an indirect buffer8. With a numeric prefix argument N, go up to level N and then take that tree. If N is negative then go up that many levels. With a C-u prefix, do not remove the previously used indirect buffer.
C-c C-x v     (org-copy-visible)
Copy the visible text in the region into the kill ring.

Next: , Previous: Global and local cycling, Up: Visibility cycling

2.3.2 Initial visibility

When Emacs first visits an Org file, the global state is set to OVERVIEW, i.e., only the top level headlines are visible9. This can be configured through the variable org-startup-folded, or on a per-file basis by adding one of the following lines anywhere in the buffer:

     #+STARTUP: overview
     #+STARTUP: content
     #+STARTUP: showall
     #+STARTUP: showeverything

The startup visibility options are ignored when the file is open for the first time during the agenda generation: if you want the agenda to honor the startup visibility, set org-agenda-inhibit-startup to nil.

Furthermore, any entries with a ‘VISIBILITY’ property (see Properties and Columns) will get their visibility adapted accordingly. Allowed values for this property are folded, children, content, and all.

C-u C-u <TAB>     (org-set-startup-visibility)
Switch back to the startup visibility of the buffer, i.e., whatever is requested by startup options and ‘VISIBILITY’ properties in individual entries.

Previous: Initial visibility, Up: Visibility cycling

2.3.3 Catching invisible edits

Sometimes you may inadvertently edit an invisible part of the buffer and be confused on what has been edited and how to undo the mistake. Setting org-catch-invisible-edits to non-nil will help prevent this. See the docstring of this option on how Org should catch invisible edits and process them.

Next: , Previous: Visibility cycling, Up: Document Structure

2.4 Motion

The following commands jump to other headlines in the buffer.

C-c C-n     (outline-next-visible-heading)
Next heading.
C-c C-p     (outline-previous-visible-heading)
Previous heading.
C-c C-f     (org-forward-same-level)
Next heading same level.
C-c C-b     (org-backward-same-level)
Previous heading same level.
C-c C-u     (outline-up-heading)
Backward to higher level heading.
C-c C-j     (org-goto)
Jump to a different place without changing the current outline visibility. Shows the document structure in a temporary buffer, where you can use the following keys to find your destination:
          <TAB>         Cycle visibility.
          <down> / <up>   Next/previous visible headline.
          <RET>         Select this location.
          /           Do a Sparse-tree search
          The following keys work if you turn off org-goto-auto-isearch
          n / p        Next/previous visible headline.
          f / b        Next/previous headline same level.
          u            One level up.
          0-9          Digit argument.
          q            Quit

See also the option org-goto-interface.

Next: , Previous: Motion, Up: Document Structure

2.5 Structure editing

M-<RET>     (org-insert-heading)
Insert a new heading/item with the same level than the one at point. If the cursor is in a plain list item, a new item is created (see Plain lists). To prevent this behavior in lists, call the command with a prefix argument. When this command is used in the middle of a line, the line is split and the rest of the line becomes the new item or headline10. If the command is used at the beginning of a headline, the new headline is created before the current line. If the command is used at the end of a folded subtree (i.e., behind the ellipses at the end of a headline), then a headline will be inserted after the end of the subtree. Calling this command with C-u C-u will unconditionally respect the headline's content and create a new item at the end of the parent subtree.
C-<RET>     (org-insert-heading-respect-content)
Just like M-<RET>, except when adding a new heading below the current heading, the new heading is placed after the body instead of before it. This command works from anywhere in the entry.
M-S-<RET>     (org-insert-todo-heading)
Insert new TODO entry with same level as current heading. See also the variable org-treat-insert-todo-heading-as-state-change.
C-S-<RET>     (org-insert-todo-heading-respect-content)
Insert new TODO entry with same level as current heading. Like C-<RET>, the new headline will be inserted after the current subtree.
<TAB>     (org-cycle)
In a new entry with no text yet, the first <TAB> demotes the entry to become a child of the previous one. The next <TAB> makes it a parent, and so on, all the way to top level. Yet another <TAB>, and you are back to the initial level.
M-<left>     (org-do-promote)
Promote current heading by one level.
M-<right>     (org-do-demote)
Demote current heading by one level.
M-S-<left>     (org-promote-subtree)
Promote the current subtree by one level.
M-S-<right>     (org-demote-subtree)
Demote the current subtree by one level.
M-S-<up>     (org-move-subtree-up)
Move subtree up (swap with previous subtree of same level).
M-S-<down>     (org-move-subtree-down)
Move subtree down (swap with next subtree of same level).
M-h     (org-mark-element)
Mark the element at point. Hitting repeatedly will mark subsequent elements of the one just marked. E.g., hitting <M-h> on a paragraph will mark it, hitting <M-h> immediately again will mark the next one.
C-c @     (org-mark-subtree)
Mark the subtree at point. Hitting repeatedly will mark subsequent subtrees of the same level than the marked subtree.
C-c C-x C-w     (org-cut-subtree)
Kill subtree, i.e., remove it from buffer but save in kill ring. With a numeric prefix argument N, kill N sequential subtrees.
C-c C-x M-w     (org-copy-subtree)
Copy subtree to kill ring. With a numeric prefix argument N, copy the N sequential subtrees.
C-c C-x C-y     (org-paste-subtree)
Yank subtree from kill ring. This does modify the level of the subtree to make sure the tree fits in nicely at the yank position. The yank level can also be specified with a numeric prefix argument, or by yanking after a headline marker like ‘****’.
C-y     (org-yank)
Depending on the options org-yank-adjusted-subtrees and org-yank-folded-subtrees, Org's internal yank command will paste subtrees folded and in a clever way, using the same command as C-c C-x C-y. With the default settings, no level adjustment will take place, but the yanked tree will be folded unless doing so would swallow text previously visible. Any prefix argument to this command will force a normal yank to be executed, with the prefix passed along. A good way to force a normal yank is C-u C-y. If you use yank-pop after a yank, it will yank previous kill items plainly, without adjustment and folding.
C-c C-x c     (org-clone-subtree-with-time-shift)
Clone a subtree by making a number of sibling copies of it. You will be prompted for the number of copies to make, and you can also specify if any timestamps in the entry should be shifted. This can be useful, for example, to create a number of tasks related to a series of lectures to prepare. For more details, see the docstring of the command org-clone-subtree-with-time-shift.
C-c C-w     (org-refile)
Refile entry or region to a different location. See Refile and copy.
C-c ^     (org-sort)
Sort same-level entries. When there is an active region, all entries in the region will be sorted. Otherwise the children of the current headline are sorted. The command prompts for the sorting method, which can be alphabetically, numerically, by time (first timestamp with active preferred, creation time, scheduled time, deadline time), by priority, by TODO keyword (in the sequence the keywords have been defined in the setup) or by the value of a property. Reverse sorting is possible as well. You can also supply your own function to extract the sorting key. With a C-u prefix, sorting will be case-sensitive.
C-x n s     (org-narrow-to-subtree)
Narrow buffer to current subtree.
C-x n b     (org-narrow-to-block)
Narrow buffer to current block.
C-x n w     (widen)
Widen buffer to remove narrowing.
C-c *     (org-toggle-heading)
Turn a normal line or plain list item into a headline (so that it becomes a subheading at its location). Also turn a headline into a normal line by removing the stars. If there is an active region, turn all lines in the region into headlines. If the first line in the region was an item, turn only the item lines into headlines. Finally, if the first line is a headline, remove the stars from all headlines in the region.

When there is an active region (Transient Mark mode), promotion and demotion work on all headlines in the region. To select a region of headlines, it is best to place both point and mark at the beginning of a line, mark at the beginning of the first headline, and point at the line just after the last headline to change. Note that when the cursor is inside a table (see Tables), the Meta-Cursor keys have different functionality.

Next: , Previous: Structure editing, Up: Document Structure

2.6 Sparse trees

An important feature of Org mode is the ability to construct sparse trees for selected information in an outline tree, so that the entire document is folded as much as possible, but the selected information is made visible along with the headline structure above it11. Just try it out and you will see immediately how it works.

Org mode contains several commands creating such trees, all these commands can be accessed through a dispatcher:

C-c /     (org-sparse-tree)
This prompts for an extra key to select a sparse-tree creating command.
C-c / r     (org-occur)
Prompts for a regexp and shows a sparse tree with all matches. If the match is in a headline, the headline is made visible. If the match is in the body of an entry, headline and body are made visible. In order to provide minimal context, also the full hierarchy of headlines above the match is shown, as well as the headline following the match. Each match is also highlighted; the highlights disappear when the buffer is changed by an editing command12, or by pressing C-c C-c. When called with a C-u prefix argument, previous highlights are kept, so several calls to this command can be stacked.
M-g n or M-g M-n     (next-error)
Jump to the next sparse tree match in this buffer.
M-g p or M-g M-p     (previous-error)
Jump to the previous sparse tree match in this buffer.

For frequently used sparse trees of specific search strings, you can use the option org-agenda-custom-commands to define fast keyboard access to specific sparse trees. These commands will then be accessible through the agenda dispatcher (see Agenda dispatcher). For example:

     (setq org-agenda-custom-commands
           '(("f" occur-tree "FIXME")))

will define the key C-c a f as a shortcut for creating a sparse tree matching the string ‘FIXME’.

The other sparse tree commands select headings based on TODO keywords, tags, or properties and will be discussed later in this manual.

To print a sparse tree, you can use the Emacs command ps-print-buffer-with-faces which does not print invisible parts of the document 13. Or you can use C-c C-e C-v to export only the visible part of the document and print the resulting file.

Next: , Previous: Sparse trees, Up: Document Structure

2.7 Plain lists

Within an entry of the outline tree, hand-formatted lists can provide additional structure. They also provide a way to create lists of checkboxes (see Checkboxes). Org supports editing such lists, and every exporter (see Exporting) can parse and format them.

Org knows ordered lists, unordered lists, and description lists.

Items belonging to the same list must have the same indentation on the first line. In particular, if an ordered list reaches number ‘10.’, then the 2–digit numbers must be written left-aligned with the other numbers in the list. An item ends before the next line that is less or equally indented than its bullet/number.

A list ends whenever every item has ended, which means before any line less or equally indented than items at top level. It also ends before two blank lines18. In that case, all items are closed. Here is an example:

     ** Lord of the Rings
        My favorite scenes are (in this order)
        1. The attack of the Rohirrim
        2. Eowyn's fight with the witch king
           + this was already my favorite scene in the book
           + I really like Miranda Otto.
        3. Peter Jackson being shot by Legolas
           - on DVD only
           He makes a really funny face when it happens.
        But in the end, no individual scenes matter but the film as a whole.
        Important actors in this film are:
        - Elijah Wood :: He plays Frodo
        - Sean Austin :: He plays Sam, Frodo's friend.  I still remember
          him very well from his role as Mikey Walsh in The Goonies.

Org supports these lists by tuning filling and wrapping commands to deal with them correctly19, and by exporting them properly (see Exporting). Since indentation is what governs the structure of these lists, many structural constructs like #+BEGIN_... blocks can be indented to signal that they belong to a particular item.

If you find that using a different bullet for a sub-list (than that used for the current list-level) improves readability, customize the variable org-list-demote-modify-bullet. To get a greater difference of indentation between items and theirs sub-items, customize org-list-indent-offset.

The following commands act on items when the cursor is in the first line of an item (the line with the bullet or number). Some of them imply the application of automatic rules to keep list structure intact. If some of these actions get in your way, configure org-list-automatic-rules to disable them individually.

<TAB>     (org-cycle)
Items can be folded just like headline levels. Normally this works only if the cursor is on a plain list item. For more details, see the variable org-cycle-include-plain-lists. If this variable is set to integrate, plain list items will be treated like low-level headlines. The level of an item is then given by the indentation of the bullet/number. Items are always subordinate to real headlines, however; the hierarchies remain completely separated. In a new item with no text yet, the first <TAB> demotes the item to become a child of the previous one. Subsequent <TAB>s move the item to meaningful levels in the list and eventually get it back to its initial position.
M-<RET>     (org-insert-heading)
Insert new item at current level. With a prefix argument, force a new heading (see Structure editing). If this command is used in the middle of an item, that item is split in two, and the second part becomes the new item20. If this command is executed before item's body, the new item is created before the current one.
Insert a new item with a checkbox (see Checkboxes).
Jump to the previous/next item in the current list21, but only if org-support-shift-select is off. If not, you can still use paragraph jumping commands like C-<up> and C-<down> to quite similar effect.
Move the item including subitems up/down22 (swap with previous/next item of same indentation). If the list is ordered, renumbering is automatic.
Decrease/increase the indentation of an item, leaving children alone.
Decrease/increase the indentation of the item, including subitems. Initially, the item tree is selected based on current indentation. When these commands are executed several times in direct succession, the initially selected region is used, even if the new indentation would imply a different hierarchy. To use the new hierarchy, break the command chain with a cursor motion or so.

As a special case, using this command on the very first item of a list will move the whole list. This behavior can be disabled by configuring org-list-automatic-rules. The global indentation of a list has no influence on the text after the list.

C-c C-c
If there is a checkbox (see Checkboxes) in the item line, toggle the state of the checkbox. In any case, verify bullets and indentation consistency in the whole list.
C-c -
Cycle the entire list level through the different itemize/enumerate bullets (‘-’, ‘+’, ‘*’, ‘1.’, ‘1)’) or a subset of them, depending on org-plain-list-ordered-item-terminator, the type of list, and its indentation. With a numeric prefix argument N, select the Nth bullet from this list. If there is an active region when calling this, selected text will be changed into an item. With a prefix argument, all lines will be converted to list items. If the first line already was a list item, any item marker will be removed from the list. Finally, even without an active region, a normal line will be converted into a list item.
C-c *
Turn a plain list item into a headline (so that it becomes a subheading at its location). See Structure editing, for a detailed explanation.
C-c C-*
Turn the whole plain list into a subtree of the current heading. Checkboxes (see Checkboxes) will become TODO (resp. DONE) keywords when unchecked (resp. checked).
This command also cycles bullet styles when the cursor in on the bullet or anywhere in an item line, details depending on org-support-shift-select.
C-c ^
Sort the plain list. You will be prompted for the sorting method: numerically, alphabetically, by time, by checked status for check lists, or by a custom function.

Next: , Previous: Plain lists, Up: Document Structure

2.8 Drawers

Sometimes you want to keep information associated with an entry, but you normally don't want to see it. For this, Org mode has drawers. Drawers need to be configured with the option org-drawers23. Drawers look like this:

     ** This is a headline
        Still outside the drawer
        This is inside the drawer.
        After the drawer.

You can interactively insert drawers at point by calling org-insert-drawer, which is bound to <C-c C-x d>. With an active region, this command will put the region inside the drawer. With a prefix argument, this command calls org-insert-property-drawer and add a property drawer right below the current headline. Completion over drawer keywords is also possible using <M-TAB>.

Visibility cycling (see Visibility cycling) on the headline will hide and show the entry, but keep the drawer collapsed to a single line. In order to look inside the drawer, you need to move the cursor to the drawer line and press <TAB> there. Org mode uses the PROPERTIES drawer for storing properties (see Properties and Columns), and you can also arrange for state change notes (see Tracking TODO state changes) and clock times (see Clocking work time) to be stored in a drawer LOGBOOK. If you want to store a quick note in the LOGBOOK drawer, in a similar way to state changes, use

C-c C-z
Add a time-stamped note to the LOGBOOK drawer.

You can select the name of the drawers which should be exported with org-export-with-drawers. In that case, drawer contents will appear in export output. Property drawers are not affected by this variable and are never exported.

Next: , Previous: Drawers, Up: Document Structure

2.9 Blocks

Org mode uses begin...end blocks for various purposes from including source code examples (see Literal examples) to capturing time logging information (see Clocking work time). These blocks can be folded and unfolded by pressing TAB in the begin line. You can also get all blocks folded at startup by configuring the option org-hide-block-startup or on a per-file basis by using

     #+STARTUP: hideblocks
     #+STARTUP: nohideblocks

Next: , Previous: Blocks, Up: Document Structure

2.10 Footnotes

Org mode supports the creation of footnotes. In contrast to the footnote.el package, Org mode's footnotes are designed for work on a larger document, not only for one-off documents like emails.

A footnote is started by a footnote marker in square brackets in column 0, no indentation allowed. It ends at the next footnote definition, headline, or after two consecutive empty lines. The footnote reference is simply the marker in square brackets, inside text. For example:

     The Org homepage[fn:1] now looks a lot better than it used to.
     [fn:1] The link is:

Org mode extends the number-based syntax to named footnotes and optional inline definition. Using plain numbers as markers (as footnote.el does) is supported for backward compatibility, but not encouraged because of possible conflicts with LaTeX snippets (see Embedded LaTeX). Here are the valid references:

A plain numeric footnote marker. Compatible with footnote.el, but not recommended because something like ‘[1]’ could easily be part of a code snippet.
A named footnote reference, where name is a unique label word, or, for simplicity of automatic creation, a number.
[fn:: This is the inline definition of this footnote]
A LaTeX-like anonymous footnote where the definition is given directly at the reference point.
[fn:name: a definition]
An inline definition of a footnote, which also specifies a name for the note. Since Org allows multiple references to the same note, you can then use [fn:name] to create additional references.

Footnote labels can be created automatically, or you can create names yourself. This is handled by the variable org-footnote-auto-label and its corresponding #+STARTUP keywords. See the docstring of that variable for details.

The following command handles footnotes:

C-c C-x f
The footnote action command.

When the cursor is on a footnote reference, jump to the definition. When it is at a definition, jump to the (first) reference.

Otherwise, create a new footnote. Depending on the option org-footnote-define-inline24, the definition will be placed right into the text as part of the reference, or separately into the location determined by the option org-footnote-section.

When this command is called with a prefix argument, a menu of additional options is offered:

          s   Sort the footnote definitions by reference sequence.  During editing,
              Org makes no effort to sort footnote definitions into a particular
              sequence.  If you want them sorted, use this command, which will
              also move entries according to org-footnote-section.  Automatic
              sorting after each insertion/deletion can be configured using the
              option org-footnote-auto-adjust.
          r   Renumber the simple fn:N footnotes.  Automatic renumbering
              after each insertion/deletion can be configured using the option
          S   Short for first r, then s action.
          n   Normalize the footnotes by collecting all definitions (including
              inline definitions) into a special section, and then numbering them
              in sequence.  The references will then also be numbers.  This is
              meant to be the final step before finishing a document (e.g., sending
              off an email).
          d   Delete the footnote at point, and all definitions of and references
              to it.

Depending on the variable org-footnote-auto-adjust25, renumbering and sorting footnotes can be automatic after each insertion or deletion.

C-c C-c
If the cursor is on a footnote reference, jump to the definition. If it is a the definition, jump back to the reference. When called at a footnote location with a prefix argument, offer the same menu as C-c C-x f.
C-c C-o or mouse-1/2
Footnote labels are also links to the corresponding definition/reference, and you can use the usual commands to follow these links.

Next: , Previous: Footnotes, Up: Document Structure

2.11 The Orgstruct minor mode

If you like the intuitive way the Org mode structure editing and list formatting works, you might want to use these commands in other modes like Text mode or Mail mode as well. The minor mode orgstruct-mode makes this possible. Toggle the mode with M-x orgstruct-mode RET, or turn it on by default, for example in Message mode, with one of:

     (add-hook 'message-mode-hook 'turn-on-orgstruct)
     (add-hook 'message-mode-hook 'turn-on-orgstruct++)

When this mode is active and the cursor is on a line that looks to Org like a headline or the first line of a list item, most structure editing commands will work, even if the same keys normally have different functionality in the major mode you are using. If the cursor is not in one of those special lines, Orgstruct mode lurks silently in the shadows.

When you use orgstruct++-mode, Org will also export indentation and autofill settings into that mode, and detect item context after the first line of an item.

You can also use Org structure editing to fold and unfold headlines in any file, provided you defined orgstruct-heading-prefix-regexp: the regular expression must match the local prefix to use before Org's headlines. For example, if you set this variable to ";; " in Emacs Lisp files, you will be able to fold and unfold headlines in Emacs Lisp commented lines. Some commands like org-demote are disabled when the prefix is set, but folding/unfolding will work correctly.

Previous: Orgstruct mode, Up: Document Structure

2.12 Org syntax

A reference document providing a formal description of Org's syntax is available as a draft on Worg, written and maintained by Nicolas Goaziou. It defines Org's core internal concepts such as headlines, sections, affiliated keywords, (greater) elements and objects. Each part of an Org file falls into one of the categories above.

To explore the abstract structure of an Org buffer, run this in a buffer:

     M-: (org-element-parse-buffer) RET

It will output a list containing the buffer's content represented as an abstract structure. The export engine relies on the information stored in this list. Most interactive commands (e.g., for structure editing) also rely on the syntactic meaning of the surrounding context.

Next: , Previous: Document Structure, Up: Top

3 Tables

Org comes with a fast and intuitive table editor. Spreadsheet-like calculations are supported using the Emacs calc package (see Calc).

Next: , Previous: Tables, Up: Tables

3.1 The built-in table editor

Org makes it easy to format tables in plain ASCII. Any line with ‘|’ as the first non-whitespace character is considered part of a table. ‘|’ is also the column separator26. A table might look like this:

     | Name  | Phone | Age |
     | Peter |  1234 |  17 |
     | Anna  |  4321 |  25 |

A table is re-aligned automatically each time you press <TAB> or <RET> or C-c C-c inside the table. <TAB> also moves to the next field (<RET> to the next row) and creates new table rows at the end of the table or before horizontal lines. The indentation of the table is set by the first line. Any line starting with ‘|-’ is considered as a horizontal separator line and will be expanded on the next re-align to span the whole table width. So, to create the above table, you would only type


and then press <TAB> to align the table and start filling in fields. Even faster would be to type |Name|Phone|Age followed by C-c <RET>.

When typing text into a field, Org treats <DEL>, <Backspace>, and all character keys in a special way, so that inserting and deleting avoids shifting other fields. Also, when typing immediately after the cursor was moved into a new field with <TAB>, S-<TAB> or <RET>, the field is automatically made blank. If this behavior is too unpredictable for you, configure the options org-enable-table-editor and org-table-auto-blank-field.

Creation and conversion

C-c |     (org-table-create-or-convert-from-region)
Convert the active region to table. If every line contains at least one TAB character, the function assumes that the material is tab separated. If every line contains a comma, comma-separated values (CSV) are assumed. If not, lines are split at whitespace into fields. You can use a prefix argument to force a specific separator: C-u forces CSV, C-u C-u forces TAB, and a numeric argument N indicates that at least N consecutive spaces, or alternatively a TAB will be the separator.
If there is no active region, this command creates an empty Org table. But it is easier just to start typing, like |Name|Phone|Age <RET> |- <TAB>.
Re-aligning and field motion

C-c C-c     (org-table-align)
Re-align the table and don't move to another field.
<TAB>     (org-table-next-field)
Re-align the table, move to the next field. Creates a new row if necessary.
S-<TAB>     (org-table-previous-field)
Re-align, move to previous field.
<RET>     (org-table-next-row)
Re-align the table and move down to next row. Creates a new row if necessary. At the beginning or end of a line, <RET> still does NEWLINE, so it can be used to split a table.
M-a     (org-table-beginning-of-field)
Move to beginning of the current table field, or on to the previous field.
M-e     (org-table-end-of-field)
Move to end of the current table field, or on to the next field.
Column and row editing

M-<left>     (org-table-move-column-left)
M-<right>     (org-table-move-column-right)
Move the current column left/right.
M-S-<left>     (org-table-delete-column)
Kill the current column.
M-S-<right>     (org-table-insert-column)
Insert a new column to the left of the cursor position.
M-<up>     (org-table-move-row-up)
M-<down>     (org-table-move-row-down)
Move the current row up/down.
M-S-<up>     (org-table-kill-row)
Kill the current row or horizontal line.
M-S-<down>     (org-table-insert-row)
Insert a new row above the current row. With a prefix argument, the line is created below the current one.
C-c -     (org-table-insert-hline)
Insert a horizontal line below current row. With a prefix argument, the line is created above the current line.
C-c <RET>     (org-table-hline-and-move)
Insert a horizontal line below current row, and move the cursor into the row below that line.
C-c ^     (org-table-sort-lines)
Sort the table lines in the region. The position of point indicates the column to be used for sorting, and the range of lines is the range between the nearest horizontal separator lines, or the entire table. If point is before the first column, you will be prompted for the sorting column. If there is an active region, the mark specifies the first line and the sorting column, while point should be in the last line to be included into the sorting. The command prompts for the sorting type (alphabetically, numerically, or by time). When called with a prefix argument, alphabetic sorting will be case-sensitive.

C-c C-x M-w     (org-table-copy-region)
Copy a rectangular region from a table to a special clipboard. Point and mark determine edge fields of the rectangle. If there is no active region, copy just the current field. The process ignores horizontal separator lines.
C-c C-x C-w     (org-table-cut-region)
Copy a rectangular region from a table to a special clipboard, and blank all fields in the rectangle. So this is the “cut” operation.
C-c C-x C-y     (org-table-paste-rectangle)
Paste a rectangular region into a table. The upper left corner ends up in the current field. All involved fields will be overwritten. If the rectangle does not fit into the present table, the table is enlarged as needed. The process ignores horizontal separator lines.
M-<RET>     (org-table-wrap-region)
Split the current field at the cursor position and move the rest to the line below. If there is an active region, and both point and mark are in the same column, the text in the column is wrapped to minimum width for the given number of lines. A numeric prefix argument may be used to change the number of desired lines. If there is no region, but you specify a prefix argument, the current field is made blank, and the content is appended to the field above.

C-c +     (org-table-sum)
Sum the numbers in the current column, or in the rectangle defined by the active region. The result is shown in the echo area and can be inserted with C-y.
S-<RET>     (org-table-copy-down)
When current field is empty, copy from first non-empty field above. When not empty, copy current field down to next row and move cursor along with it. Depending on the option org-table-copy-increment, integer field values will be incremented during copy. Integers that are too large will not be incremented. Also, a 0 prefix argument temporarily disables the increment. This key is also used by shift-selection and related modes (see Conflicts).

C-c `     (org-table-edit-field)
Edit the current field in a separate window. This is useful for fields that are not fully visible (see Column width and alignment). When called with a C-u prefix, just make the full field visible, so that it can be edited in place. When called with two C-u prefixes, make the editor window follow the cursor through the table and always show the current field. The follow mode exits automatically when the cursor leaves the table, or when you repeat this command with C-u C-u C-c `.
M-x org-table-import RET
Import a file as a table. The table should be TAB or whitespace separated. Use, for example, to import a spreadsheet table or data from a database, because these programs generally can write TAB-separated text files. This command works by inserting the file into the buffer and then converting the region to a table. Any prefix argument is passed on to the converter, which uses it to determine the separator.
C-c |     (org-table-create-or-convert-from-region)
Tables can also be imported by pasting tabular text into the Org buffer, selecting the pasted text with C-x C-x and then using the C-c | command (see above under Creation and conversion).
M-x org-table-export RET
Export the table, by default as a TAB-separated file. Use for data exchange with, for example, spreadsheet or database programs. The format used to export the file can be configured in the option org-table-export-default-format. You may also use properties TABLE_EXPORT_FILE and TABLE_EXPORT_FORMAT to specify the file name and the format for table export in a subtree. Org supports quite general formats for exported tables. The exporter format is the same as the format used by Orgtbl radio tables, see Translator functions, for a detailed description.

If you don't like the automatic table editor because it gets in your way on lines which you would like to start with ‘|’, you can turn it off with

     (setq org-enable-table-editor nil)

Then the only table command that still works is C-c C-c to do a manual re-align.

Next: , Previous: Built-in table editor, Up: Tables

3.2 Column width and alignment

The width of columns is automatically determined by the table editor. And also the alignment of a column is determined automatically from the fraction of number-like versus non-number fields in the column.

Sometimes a single field or a few fields need to carry more text, leading to inconveniently wide columns. Or maybe you want to make a table with several columns having a fixed width, regardless of content. To set27 the width of a column, one field anywhere in the column may contain just the string ‘<N>’ where ‘N’ is an integer specifying the width of the column in characters. The next re-align will then set the width of this column to this value.

     |---+------------------------------|               |---+--------|
     |   |                              |               |   | <6>    |
     | 1 | one                          |               | 1 | one    |
     | 2 | two                          |     ----\     | 2 | two    |
     | 3 | This is a long chunk of text |     ----/     | 3 | This=> |
     | 4 | four                         |               | 4 | four   |
     |---+------------------------------|               |---+--------|

Fields that are wider become clipped and end in the string ‘=>’. Note that the full text is still in the buffer but is hidden. To see the full text, hold the mouse over the field—a tool-tip window will show the full content. To edit such a field, use the command C-c ` (that is C-c followed by the grave accent). This will open a new window with the full field. Edit it and finish with C-c C-c.

When visiting a file containing a table with narrowed columns, the necessary character hiding has not yet happened, and the table needs to be aligned before it looks nice. Setting the option org-startup-align-all-tables will realign all tables in a file upon visiting, but also slow down startup. You can also set this option on a per-file basis with:

     #+STARTUP: align
     #+STARTUP: noalign

If you would like to overrule the automatic alignment of number-rich columns to the right and of string-rich column to the left, you can use ‘<r>’, ‘<c>28 or ‘<l>’ in a similar fashion. You may also combine alignment and field width like this: ‘<r10>’.

Lines which only contain these formatting cookies will be removed automatically when exporting the document.

Next: , Previous: Column width and alignment, Up: Tables

3.3 Column groups

When Org exports tables, it does so by default without vertical lines because that is visually more satisfying in general. Occasionally however, vertical lines can be useful to structure a table into groups of columns, much like horizontal lines can do for groups of rows. In order to specify column groups, you can use a special row where the first field contains only ‘/’. The further fields can either contain ‘<’ to indicate that this column should start a group, ‘>’ to indicate the end of a column, or ‘<>’ (no space between ‘<’ and ‘>’) to make a column a group of its own. Boundaries between column groups will upon export be marked with vertical lines. Here is an example:

     | N | N^2 | N^3 | N^4 | sqrt(n) | sqrt[4](N) |
     | / |   < |     |   > |       < |          > |
     | 1 |   1 |   1 |   1 |       1 |          1 |
     | 2 |   4 |   8 |  16 |  1.4142 |     1.1892 |
     | 3 |   9 |  27 |  81 |  1.7321 |     1.3161 |
     #+TBLFM: $2=$1^2::$3=$1^3::$4=$1^4::$5=sqrt($1)::$6=sqrt(sqrt(($1)))

It is also sufficient to just insert the column group starters after every vertical line you would like to have:

     |  N | N^2 | N^3 | N^4 | sqrt(n) | sqrt[4](N) |
     | /  | <   |     |     | <       |            |

Next: , Previous: Column groups, Up: Tables

3.4 The Orgtbl minor mode

If you like the intuitive way the Org table editor works, you might also want to use it in other modes like Text mode or Mail mode. The minor mode Orgtbl mode makes this possible. You can always toggle the mode with M-x orgtbl-mode RET. To turn it on by default, for example in Message mode, use

     (add-hook 'message-mode-hook 'turn-on-orgtbl)

Furthermore, with some special setup, it is possible to maintain tables in arbitrary syntax with Orgtbl mode. For example, it is possible to construct LaTeX tables with the underlying ease and power of Orgtbl mode, including spreadsheet capabilities. For details, see Tables in arbitrary syntax.

Next: , Previous: Orgtbl mode, Up: Tables

3.5 The spreadsheet

The table editor makes use of the Emacs calc package to implement spreadsheet-like capabilities. It can also evaluate Emacs Lisp forms to derive fields from other fields. While fully featured, Org's implementation is not identical to other spreadsheets. For example, Org knows the concept of a column formula that will be applied to all non-header fields in a column without having to copy the formula to each relevant field. There is also a formula debugger, and a formula editor with features for highlighting fields in the table corresponding to the references at the point in the formula, moving these references by arrow keys

Next: , Previous: The spreadsheet, Up: The spreadsheet

3.5.1 References

To compute fields in the table from other fields, formulas must reference other fields or ranges. In Org, fields can be referenced by name, by absolute coordinates, and by relative coordinates. To find out what the coordinates of a field are, press C-c ? in that field, or press C-c } to toggle the display of a grid.

Field references

Formulas can reference the value of another field in two ways. Like in any other spreadsheet, you may reference fields with a letter/number combination like B3, meaning the 2nd field in the 3rd row. However, Org prefers29 to use another, more general representation that looks like this:


Column specifications can be absolute like $1, $2,...$N, or relative to the current column (i.e., the column of the field which is being computed) like $+1 or $-2. $< and $> are immutable references to the first and last column, respectively, and you can use $>>> to indicate the third column from the right.

The row specification only counts data lines and ignores horizontal separator lines (hlines). Like with columns, you can use absolute row numbers @1, @2,...@N, and row numbers relative to the current row like @+3 or @-1. @< and @> are immutable references the first and last30 row in the table, respectively. You may also specify the row relative to one of the hlines: @I refers to the first hline, @II to the second, etc. @-I refers to the first such line above the current line, @+I to the first such line below the current line. You can also write @III+2 which is the second data line after the third hline in the table.

@0 and $0 refer to the current row and column, respectively, i.e., to the row/column for the field being computed. Also, if you omit either the column or the row part of the reference, the current row/column is implied.

Org's references with unsigned numbers are fixed references in the sense that if you use the same reference in the formula for two different fields, the same field will be referenced each time. Org's references with signed numbers are floating references because the same reference operator can reference different fields depending on the field being calculated by the formula.

Here are a few examples:

     @2$3      2nd row, 3rd column (same as C2)
     $5        column 5 in the current row (same as E&)
     @2        current column, row 2
     @-1$-3    the field one row up, three columns to the left
     @-I$2     field just under hline above current row, column 2
     @>$5      field in the last row, in column 5
Range references

You may reference a rectangular range of fields by specifying two field references connected by two dots ‘..’. If both fields are in the current row, you may simply use ‘$2..$7’, but if at least one field is in a different row, you need to use the general @row$column format at least for the first field (i.e the reference must start with ‘@’ in order to be interpreted correctly). Examples:

     $1..$3        first three fields in the current row
     $P..$Q        range, using column names (see under Advanced)
     $<<<..$>>     start in third column, continue to the one but last
     @2$1..@4$3    6 fields between these two fields (same as A2..C4)
     @-1$-2..@-1   3 fields in the row above, starting from 2 columns on the left
     @I..II        between first and second hline, short for @I..@II

Range references return a vector of values that can be fed into Calc vector functions. Empty fields in ranges are normally suppressed, so that the vector contains only the non-empty fields. For other options with the mode switches ‘E’, ‘N’ and examples see Formula syntax for Calc.

Field coordinates in formulas

For Calc formulas and Lisp formulas @# and $# can be used to get the row or column number of the field where the formula result goes. The traditional Lisp formula equivalents are org-table-current-dline and org-table-current-column. Examples:

     if(@# % 2, $#, string(""))   column number on odd lines only
     $3 = remote(FOO, @@#$2)      copy column 2 from table FOO into
                                  column 3 of the current table

For the second example, table FOO must have at least as many rows as the current table. Note that this is inefficient31 for large number of rows.

Named references

$name’ is interpreted as the name of a column, parameter or constant. Constants are defined globally through the option org-table-formula-constants, and locally (for the file) through a line like

     #+CONSTANTS: c=299792458. pi=3.14 eps=2.4e-6

Also properties (see Properties and Columns) can be used as constants in table formulas: for a property ‘:Xyz:’ use the name ‘$PROP_Xyz’, and the property will be searched in the current outline entry and in the hierarchy above it. If you have the constants.el package, it will also be used to resolve constants, including natural constants like ‘$h’ for Planck's constant, and units like ‘$km’ for kilometers32. Column names and parameters can be specified in special table lines. These are described below, see Advanced features. All names must start with a letter, and further consist of letters and numbers.

Remote references

You may also reference constants, fields and ranges from a different table, either in the current file or even in a different file. The syntax is


where NAME can be the name of a table in the current file as set by a #+NAME: Name line before the table. It can also be the ID of an entry, even in a different file, and the reference then refers to the first table in that entry. REF is an absolute field or range reference as described above for example @3$3 or $somename, valid in the referenced table.

Next: , Previous: References, Up: The spreadsheet

3.5.2 Formula syntax for Calc

A formula can be any algebraic expression understood by the Emacs Calc package. Note that calc has the non-standard convention that ‘/’ has lower precedence than ‘*’, so that ‘a/b*c’ is interpreted as ‘a/(b*c)’. Before evaluation by calc-eval (see calc-eval), variable substitution takes place according to the rules described above. The range vectors can be directly fed into the Calc vector functions like ‘vmean’ and ‘vsum’.

A formula can contain an optional mode string after a semicolon. This string consists of flags to influence Calc and other modes during execution. By default, Org uses the standard Calc modes (precision 12, angular units degrees, fraction and symbolic modes off). The display format, however, has been changed to (float 8) to keep tables compact. The default settings can be configured using the option org-calc-default-modes.

List of modes:

Set the internal Calc calculation precision to 20 digits.
n3, s3, e2, f4
Normal, scientific, engineering or fixed format of the result of Calc passed back to Org. Calc formatting is unlimited in precision as long as the Calc calculation precision is greater.
D, R
Degree and radian angle modes of Calc.
F, S
Fraction and symbolic modes of Calc.
T, t
Duration computations in Calc or Lisp, see Durations and time values.
If and how to consider empty fields. Without ‘E’ empty fields in range references are suppressed so that the Calc vector or Lisp list contains only the non-empty fields. With ‘E’ the empty fields are kept. For empty fields in ranges or empty field references the value ‘nan’ (not a number) is used in Calc formulas and the empty string is used for Lisp formulas. Add ‘N’ to use 0 instead for both formula types. For the value of a field the mode ‘N’ has higher precedence than ‘E’.
Interpret all fields as numbers, use 0 for non-numbers. See the next section to see how this is essential for computations with Lisp formulas. In Calc formulas it is used only occasionally because there number strings are already interpreted as numbers without ‘N’.
Literal, for Lisp formulas only. See the next section.

Unless you use large integer numbers or high-precision-calculation and -display for floating point numbers you may alternatively provide a ‘printf’ format specifier to reformat the Calc result after it has been passed back to Org instead of letting Calc already do the formatting33. A few examples:

     $1+$2                Sum of first and second field
     $1+$2;%.2f           Same, format result to two decimals
     exp($2)+exp($1)      Math functions can be used
     $0;%.1f              Reformat current cell to 1 decimal
     ($3-32)*5/9          Degrees F -> C conversion
     $c/$1/$cm            Hz -> cm conversion, using constants.el
     tan($1);Dp3s1        Compute in degrees, precision 3, display SCI 1
     sin($1);Dp3%.1e      Same, but use printf specifier for display
     taylor($3,x=7,2)     Taylor series of $3, at x=7, second degree

Calc also contains a complete set of logical operations, (see Logical Operations). For example

if($1 < 20, teen, string(""))
"teen" if age $1 is less than 20, else the Org table result field is set to empty with the empty string.
if("$1" == "nan" || "$2" == "nan", string(""), $1 + $2); E f-1
Sum of the first two columns. When at least one of the input fields is empty the Org table result field is set to empty. ‘E’ is required to not convert empty fields to 0. ‘f-1’ is an optional Calc format string similar to ‘%.1f’ but leaves empty results empty.
if(typeof(vmean($1..$7)) == 12, string(""), vmean($1..$7); E
Mean value of a range unless there is any empty field. Every field in the range that is empty is replaced by ‘nan’ which lets ‘vmean’ result in ‘nan’. Then ‘typeof == 12’ detects the ‘nan’ from ‘vmean’ and the Org table result field is set to empty. Use this when the sample set is expected to never have missing values.
if("$1..$7" == "[]", string(""), vmean($1..$7))
Mean value of a range with empty fields skipped. Every field in the range that is empty is skipped. When all fields in the range are empty the mean value is not defined and the Org table result field is set to empty. Use this when the sample set can have a variable size.
vmean($1..$7); EN
To complete the example before: Mean value of a range with empty fields counting as samples with value 0. Use this only when incomplete sample sets should be padded with 0 to the full size.

You can add your own Calc functions defined in Emacs Lisp with defmath and use them in formula syntax for Calc.

Next: , Previous: Formula syntax for Calc, Up: The spreadsheet

3.5.3 Emacs Lisp forms as formulas

It is also possible to write a formula in Emacs Lisp. This can be useful for string manipulation and control structures, if Calc's functionality is not enough.

If a formula starts with an apostrophe followed by an opening parenthesis, then it is evaluated as a Lisp form. The evaluation should return either a string or a number. Just as with calc formulas, you can specify modes and a printf format after a semicolon.

With Emacs Lisp forms, you need to be conscious about the way field references are interpolated into the form. By default, a reference will be interpolated as a Lisp string (in double-quotes) containing the field. If you provide the ‘N’ mode switch, all referenced elements will be numbers (non-number fields will be zero) and interpolated as Lisp numbers, without quotes. If you provide the ‘L’ flag, all fields will be interpolated literally, without quotes. I.e., if you want a reference to be interpreted as a string by the Lisp form, enclose the reference operator itself in double-quotes, like "$3". Ranges are inserted as space-separated fields, so you can embed them in list or vector syntax.

Here are a few examples—note how the ‘N’ mode is used when we do computations in Lisp:

'(concat (substring $1 1 2) (substring $1 0 1) (substring $1 2))
Swap the first two characters of the content of column 1.
'(+ $1 $2);N
Add columns 1 and 2, equivalent to Calc's $1+$2.
'(apply '+ '($1..$4));N
Compute the sum of columns 1 to 4, like Calc's vsum($1..$4).

Next: , Previous: Formula syntax for Lisp, Up: The spreadsheet

3.5.4 Durations and time values

If you want to compute time values use the T flag, either in Calc formulas or Elisp formulas:

       |  Task 1 |   Task 2 |    Total |
       |    2:12 |     1:47 | 03:59:00 |
       | 3:02:20 | -2:07:00 |     0.92 |
       #+TBLFM: @2$3=$1+$2;T::@3$3=$1+$2;t

Input duration values must be of the form [HH:MM[:SS], where seconds are optional. With the T flag, computed durations will be displayed as HH:MM:SS (see the first formula above). With the t flag, computed durations will be displayed according to the value of the option org-table-duration-custom-format, which defaults to 'hours and will display the result as a fraction of hours (see the second formula in the example above).

Negative duration values can be manipulated as well, and integers will be considered as seconds in addition and subtraction.

Next: , Previous: Durations and time values, Up: The spreadsheet

3.5.5 Field and range formulas

To assign a formula to a particular field, type it directly into the field, preceded by ‘:=’, for example ‘:=vsum(@II..III)’. When you press <TAB> or <RET> or C-c C-c with the cursor still in the field, the formula will be stored as the formula for this field, evaluated, and the current field will be replaced with the result.

Formulas are stored in a special line starting with ‘#+TBLFM:’ directly below the table. If you type the equation in the 4th field of the 3rd data line in the table, the formula will look like ‘@3$4=$1+$2’. When inserting/deleting/swapping column and rows with the appropriate commands, absolute references (but not relative ones) in stored formulas are modified in order to still reference the same field. To avoid this from happening, in particular in range references, anchor ranges at the table borders (using @<, @>, $<, $>), or at hlines using the @I notation. Automatic adaptation of field references does of course not happen if you edit the table structure with normal editing commands—then you must fix the equations yourself.

Instead of typing an equation into the field, you may also use the following command

C-u C-c =     (org-table-eval-formula)
Install a new formula for the current field. The command prompts for a formula with default taken from the ‘#+TBLFM:’ line, applies it to the current field, and stores it.

The left-hand side of a formula can also be a special expression in order to assign the formula to a number of different fields. There is no keyboard shortcut to enter such range formulas. To add them, use the formula editor (see Editing and debugging formulas) or edit the #+TBLFM: line directly.

Column formula, valid for the entire column. This is so common that Org treats these formulas in a special way, see Column formulas.
Row formula, applies to all fields in the specified row. @>= means the last row.
Range formula, applies to all fields in the given rectangular range. This can also be used to assign a formula to some but not all fields in a row.
Named field, see Advanced features.

Next: , Previous: Field and range formulas, Up: The spreadsheet

3.5.6 Column formulas

When you assign a formula to a simple column reference like $3=, the same formula will be used in all fields of that column, with the following very convenient exceptions: (i) If the table contains horizontal separator hlines with rows above and below, everything before the first such hline is considered part of the table header and will not be modified by column formulas. Therefore a header is mandatory when you use column formulas and want to add hlines to group rows, like for example to separate a total row at the bottom from the summand rows above. (ii) Fields that already get a value from a field/range formula will be left alone by column formulas. These conditions make column formulas very easy to use.

To assign a formula to a column, type it directly into any field in the column, preceded by an equal sign, like ‘=$1+$2’. When you press <TAB> or <RET> or C-c C-c with the cursor still in the field, the formula will be stored as the formula for the current column, evaluated and the current field replaced with the result. If the field contains only ‘=’, the previously stored formula for this column is used. For each column, Org will only remember the most recently used formula. In the ‘#+TBLFM:’ line, column formulas will look like ‘$4=$1+$2’. The left-hand side of a column formula can not be the name of column, it must be the numeric column reference or $>.

Instead of typing an equation into the field, you may also use the following command:

C-c =     (org-table-eval-formula)
Install a new formula for the current column and replace current field with the result of the formula. The command prompts for a formula, with default taken from the ‘#+TBLFM’ line, applies it to the current field and stores it. With a numeric prefix argument(e.g., C-5 C-c =) the command will apply it to that many consecutive fields in the current column.

Next: , Previous: Column formulas, Up: The spreadsheet

3.5.7 Lookup functions

Org has three predefined Emacs Lisp functions for lookups in tables.

(org-lookup-first VAL S-LIST R-LIST &optional PREDICATE)
Searches for the first element S in list S-LIST for which
          (PREDICATE VAL S)

is t; returns the value from the corresponding position in list R-LIST. The default PREDICATE is equal. Note that the parameters VAL and S are passed to PREDICATE in the same order as the corresponding parameters are in the call to org-lookup-first, where VAL precedes S-LIST. If R-LIST is nil, the matching element S of S-LIST is returned.

(org-lookup-last VAL S-LIST R-LIST &optional PREDICATE)
Similar to org-lookup-first above, but searches for the last element for which PREDICATE is t.
(org-lookup-all VAL S-LIST R-LIST &optional PREDICATE)
Similar to org-lookup-first, but searches for all elements for which PREDICATE is t, and returns all corresponding values. This function can not be used by itself in a formula, because it returns a list of values. However, powerful lookups can be built when this function is combined with other Emacs Lisp functions.

If the ranges used in these functions contain empty fields, the E mode for the formula should usually be specified: otherwise empty fields will not be included in S-LIST and/or R-LIST which can, for example, result in an incorrect mapping from an element of S-LIST to the corresponding element of R-LIST.

These three functions can be used to implement associative arrays, count matching cells, rank results, group data etc. For practical examples see this tutorial on Worg.

Next: , Previous: Lookup functions, Up: The spreadsheet

3.5.8 Editing and debugging formulas

You can edit individual formulas in the minibuffer or directly in the field. Org can also prepare a special buffer with all active formulas of a table. When offering a formula for editing, Org converts references to the standard format (like B3 or D&) if possible. If you prefer to only work with the internal format (like @3$2 or $4), configure the option org-table-use-standard-references.

C-c = or C-u C-c =     (org-table-eval-formula)
Edit the formula associated with the current column/field in the minibuffer. See Column formulas, and Field and range formulas.
C-u C-u C-c =     (org-table-eval-formula)
Re-insert the active formula (either a field formula, or a column formula) into the current field, so that you can edit it directly in the field. The advantage over editing in the minibuffer is that you can use the command C-c ?.
C-c ?     (org-table-field-info)
While editing a formula in a table field, highlight the field(s) referenced by the reference at the cursor position in the formula.
C-c }
Toggle the display of row and column numbers for a table, using overlays (org-table-toggle-coordinate-overlays). These are updated each time the table is aligned; you can force it with C-c C-c.
C-c {
Toggle the formula debugger on and off (org-table-toggle-formula-debugger). See below.
C-c '     (org-table-edit-formulas)
Edit all formulas for the current table in a special buffer, where the formulas will be displayed one per line. If the current field has an active formula, the cursor in the formula editor will mark it. While inside the special buffer, Org will automatically highlight any field or range reference at the cursor position. You may edit, remove and add formulas, and use the following commands:
C-c C-c or C-x C-s     (org-table-fedit-finish)
Exit the formula editor and store the modified formulas. With C-u prefix, also apply the new formulas to the entire table.
C-c C-q     (org-table-fedit-abort)
Exit the formula editor without installing changes.
C-c C-r     (org-table-fedit-toggle-ref-type)
Toggle all references in the formula editor between standard (like B3) and internal (like @3$2).
<TAB>     (org-table-fedit-lisp-indent)
Pretty-print or indent Lisp formula at point. When in a line containing a Lisp formula, format the formula according to Emacs Lisp rules. Another <TAB> collapses the formula back again. In the open formula, <TAB> re-indents just like in Emacs Lisp mode.
M-<TAB>     (lisp-complete-symbol)
Complete Lisp symbols, just like in Emacs Lisp mode.
Shift the reference at point. For example, if the reference is B3 and you press S-<right>, it will become C3. This also works for relative references and for hline references.
M-S-<up>     (org-table-fedit-line-up)
M-S-<down>     (org-table-fedit-line-down)
Move the test line for column formulas in the Org buffer up and down.
M-<up>     (org-table-fedit-scroll-down)
M-<down>     (org-table-fedit-scroll-up)
Scroll the window displaying the table.
C-c }
Turn the coordinate grid in the table on and off.

Making a table field blank does not remove the formula associated with the field, because that is stored in a different line (the ‘#+TBLFM’ line)—during the next recalculation the field will be filled again. To remove a formula from a field, you have to give an empty reply when prompted for the formula, or to edit the ‘#+TBLFM’ line.

You may edit the ‘#+TBLFM’ directly and re-apply the changed equations with C-c C-c in that line or with the normal recalculation commands in the table.

Using multiple #+TBLFM lines

You may apply the formula temporarily. This is useful when you switch the formula. Place multiple ‘#+TBLFM’ lines right after the table, and then press C-c C-c on the formula to apply. Here is an example:

     | x | y |
     | 1 |   |
     | 2 |   |
     #+TBLFM: $2=$1*1
     #+TBLFM: $2=$1*2

Pressing C-c C-c in the line of ‘#+TBLFM: $2=$1*2’ yields:

     | x | y |
     | 1 | 2 |
     | 2 | 4 |
     #+TBLFM: $2=$1*1
     #+TBLFM: $2=$1*2

Note: If you recalculate this table (with C-u C-c *, for example), you will get the following result of applying only the first ‘#+TBLFM’ line.

     | x | y |
     | 1 | 1 |
     | 2 | 2 |
     #+TBLFM: $2=$1*1
     #+TBLFM: $2=$1*2
Debugging formulas

When the evaluation of a formula leads to an error, the field content becomes the string ‘#ERROR’. If you would like see what is going on during variable substitution and calculation in order to find a bug, turn on formula debugging in the Tbl menu and repeat the calculation, for example by pressing C-u C-u C-c = <RET> in a field. Detailed information will be displayed.

Next: , Previous: Editing and debugging formulas, Up: The spreadsheet

3.5.9 Updating the table

Recalculation of a table is normally not automatic, but needs to be triggered by a command. See Advanced features, for a way to make recalculation at least semi-automatic.

In order to recalculate a line of a table or the entire table, use the following commands:

C-c *     (org-table-recalculate)
Recalculate the current row by first applying the stored column formulas from left to right, and all field/range formulas in the current row.
C-u C-c *
C-u C-c C-c
Recompute the entire table, line by line. Any lines before the first hline are left alone, assuming that these are part of the table header.
C-u C-u C-c * or C-u C-u C-c C-c     (org-table-iterate)
Iterate the table by recomputing it until no further changes occur. This may be necessary if some computed fields use the value of other fields that are computed later in the calculation sequence.
M-x org-table-recalculate-buffer-tables RET
Recompute all tables in the current buffer.
M-x org-table-iterate-buffer-tables RET
Iterate all tables in the current buffer, in order to converge table-to-table dependencies.

Previous: Updating the table, Up: The spreadsheet

3.5.10 Advanced features

If you want the recalculation of fields to happen automatically, or if you want to be able to assign names34 to fields and columns, you need to reserve the first column of the table for special marking characters.

C-#     (org-table-rotate-recalc-marks)
Rotate the calculation mark in first column through the states ‘ ’, ‘#’, ‘*’, ‘!’, ‘$’. When there is an active region, change all marks in the region.

Here is an example of a table that collects exam results of students and makes use of these features:

     |   | Student | Prob 1 | Prob 2 | Prob 3 | Total | Note |
     | ! |         |     P1 |     P2 |     P3 |   Tot |      |
     | # | Maximum |     10 |     15 |     25 |    50 | 10.0 |
     | ^ |         |     m1 |     m2 |     m3 |    mt |      |
     | # | Peter   |     10 |      8 |     23 |    41 |  8.2 |
     | # | Sam     |      2 |      4 |      3 |     9 |  1.8 |
     |   | Average |        |        |        |  25.0 |      |
     | ^ |         |        |        |        |    at |      |
     | $ | max=50  |        |        |        |       |      |
     #+TBLFM: $6=vsum($P1..$P3)::$7=10*$Tot/$max;%.1f::$at=vmean(@-II..@-I);%.1f

Important: please note that for these special tables, recalculating the table with C-u C-c * will only affect rows that are marked ‘#’ or ‘*’, and fields that have a formula assigned to the field itself. The column formulas are not applied in rows with empty first field.

The marking characters have the following meaning:

The fields in this line define names for the columns, so that you may refer to a column as ‘$Tot’ instead of ‘$6’.
This row defines names for the fields above the row. With such a definition, any formula in the table may use ‘$m1’ to refer to the value ‘10’. Also, if you assign a formula to a names field, it will be stored as ‘$name=...’.
Similar to ‘^’, but defines names for the fields in the row below.
Fields in this row can define parameters for formulas. For example, if a field in a ‘$’ row contains ‘max=50’, then formulas in this table can refer to the value 50 using ‘$max’. Parameters work exactly like constants, only that they can be defined on a per-table basis.
Fields in this row are automatically recalculated when pressing <TAB> or <RET> or S-<TAB> in this row. Also, this row is selected for a global recalculation with C-u C-c *. Unmarked lines will be left alone by this command.
Selects this line for global recalculation with C-u C-c *, but not for automatic recalculation. Use this when automatic recalculation slows down editing too much.
Unmarked lines are exempt from recalculation with C-u C-c *. All lines that should be recalculated should be marked with ‘#’ or ‘*’.
Do not export this line. Useful for lines that contain the narrowing ‘<N>’ markers or column group markers.

Finally, just to whet your appetite for what can be done with the fantastic calc.el package, here is a table that computes the Taylor series of degree n at location x for a couple of functions.

     |   | Func        | n | x   | Result                               |
     | # | exp(x)      | 1 | x   | 1 + x                                |
     | # | exp(x)      | 2 | x   | 1 + x + x^2 / 2                      |
     | # | exp(x)      | 3 | x   | 1 + x + x^2 / 2 + x^3 / 6            |
     | # | x^2+sqrt(x) | 2 | x=0 | x*(0.5 / 0) + x^2 (2 - 0.25 / 0) / 2 |
     | # | x^2+sqrt(x) | 2 | x=1 | 2 + 2.5 x - 2.5 + 0.875 (x - 1)^2    |
     | * | tan(x)      | 3 | x   | 0.0175 x + 1.77e-6 x^3               |
     #+TBLFM: $5=taylor($2,$4,$3);n3

Previous: The spreadsheet, Up: Tables

3.6 Org-Plot

Org-Plot can produce 2D and 3D graphs of information stored in org tables using Gnuplot and gnuplot-mode To see this in action, ensure that you have both Gnuplot and Gnuplot mode installed on your system, then call org-plot/gnuplot on the following table.

     #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
     | Sede      | Max cites | H-index |
     | Chile     |    257.72 |   21.39 |
     | Leeds     |    165.77 |   19.68 |
     | Sao Paolo |     71.00 |   11.50 |
     | Stockholm |    134.19 |   14.33 |
     | Morelia   |    257.56 |   17.67 |

Notice that Org Plot is smart enough to apply the table's headers as labels. Further control over the labels, type, content, and appearance of plots can be exercised through the #+PLOT: lines preceding a table. See below for a complete list of Org-plot options. For more information and examples see the Org-plot tutorial at

Plot Options
Specify any gnuplot option to be set when graphing.
Specify the title of the plot.
Specify which column of the table to use as the x axis.
Specify the columns to graph as a Lisp style list, surrounded by parentheses and separated by spaces for example dep:(3 4) to graph the third and fourth columns (defaults to graphing all other columns aside from the ind column).
Specify whether the plot will be 2d, 3d, or grid.
Specify a with option to be inserted for every col being plotted (e.g., lines, points, boxes, impulses, etc...). Defaults to lines.
If you want to plot to a file, specify "path/to/desired/output-file".
List of labels to be used for the deps (defaults to the column headers if they exist).
Specify an entire line to be inserted in the Gnuplot script.
When plotting 3d or grid types, set this to t to graph a flat mapping rather than a 3d slope.
Specify format of Org mode timestamps as they will be parsed by Gnuplot. Defaults to ‘%Y-%m-%d-%H:%M:%S’.
If you want total control, you can specify a script file (place the file name between double-quotes) which will be used to plot. Before plotting, every instance of $datafile in the specified script will be replaced with the path to the generated data file. Note: even if you set this option, you may still want to specify the plot type, as that can impact the content of the data file.

Next: , Previous: Tables, Up: Top

4 Hyperlinks

Like HTML, Org provides links inside a file, external links to other files, Usenet articles, emails, and much more.

Next: , Previous: Hyperlinks, Up: Hyperlinks

4.1 Link format

Org will recognize plain URL-like links and activate them as clickable links. The general link format, however, looks like this:

     [[link][description]]       or alternatively           [[link]]

Once a link in the buffer is complete (all brackets present), Org will change the display so that ‘description’ is displayed instead of ‘[[link][description]]’ and ‘link’ is displayed instead of ‘[[link]]’. Links will be highlighted in the face org-link, which by default is an underlined face. You can directly edit the visible part of a link. Note that this can be either the ‘link’ part (if there is no description) or the ‘description’ part. To edit also the invisible ‘link’ part, use C-c C-l with the cursor on the link.

If you place the cursor at the beginning or just behind the end of the displayed text and press <BACKSPACE>, you will remove the (invisible) bracket at that location. This makes the link incomplete and the internals are again displayed as plain text. Inserting the missing bracket hides the link internals again. To show the internal structure of all links, use the menu entry Org->Hyperlinks->Literal links.

Next: , Previous: Link format, Up: Hyperlinks

4.2 Internal links

If the link does not look like a URL, it is considered to be internal in the current file. The most important case is a link like ‘[[#my-custom-id]]’ which will link to the entry with the CUSTOM_ID property ‘my-custom-id’. You are responsible yourself to make sure these custom IDs are unique in a file.

Links such as ‘[[My Target]]’ or ‘[[My Target][Find my target]]’ lead to a text search in the current file.

The link can be followed with C-c C-o when the cursor is on the link, or with a mouse click (see Handling links). Links to custom IDs will point to the corresponding headline. The preferred match for a text link is a dedicated target: the same string in double angular brackets, like ‘<<My Target>>’.

If no dedicated target exists, the link will then try to match the exact name of an element within the buffer. Naming is done with the #+NAME keyword, which has to be put the line before the element it refers to, as in the following example

     #+NAME: My Target
     | a  | table      |
     | of | four cells |

If none of the above succeeds, Org will search for a headline that is exactly the link text but may also include a TODO keyword and tags35.

During export, internal links will be used to mark objects and assign them a number. Marked objects will then be referenced by links pointing to them. In particular, links without a description will appear as the number assigned to the marked object36. In the following excerpt from an Org buffer

     - one item
     - <<target>>another item
     Here we refer to item [[target]].

The last sentence will appear as ‘Here we refer to item 2’ when exported.

In non-Org files, the search will look for the words in the link text. In the above example the search would be for ‘my target’.

Following a link pushes a mark onto Org's own mark ring. You can return to the previous position with C-c &. Using this command several times in direct succession goes back to positions recorded earlier.

Previous: Internal links, Up: Internal links

4.2.1 Radio targets

Org can automatically turn any occurrences of certain target names in normal text into a link. So without explicitly creating a link, the text connects to the target radioing its position. Radio targets are enclosed by triple angular brackets. For example, a target ‘<<<My Target>>>’ causes each occurrence of ‘my target’ in normal text to become activated as a link. The Org file is scanned automatically for radio targets only when the file is first loaded into Emacs. To update the target list during editing, press C-c C-c with the cursor on or at a target.

Next: , Previous: Internal links, Up: Hyperlinks

4.3 External links

Org supports links to files, websites, Usenet and email messages, BBDB database entries and links to both IRC conversations and their logs. External links are URL-like locators. They start with a short identifying string followed by a colon. There can be no space after the colon. The following list shows examples for each link type.          on the web
     doi:10.1000/182                           DOI for an electronic resource
     file:/home/dominik/images/jupiter.jpg     file, absolute path
     /home/dominik/images/jupiter.jpg          same as above
     file:papers/last.pdf                      file, relative path
     ./papers/last.pdf                         same as above
     file:/myself@some.where:papers/last.pdf   file, path on remote machine
     /myself@some.where:papers/last.pdf        same as above
     file:sometextfile::NNN                    file, jump to line number                         another Org file words             text search in Org file37*task title            heading search in Org file
     file+sys:/path/to/file                    open via OS, like double-click
     file+emacs:/path/to/file                  force opening by Emacs
     docview:papers/last.pdf::NNN              open in doc-view mode at page
     id:B7423F4D-2E8A-471B-8810-C40F074717E9   Link to heading by ID
     news:comp.emacs                           Usenet link                   Mail link
     mhe:folder                                MH-E folder link
     mhe:folder#id                             MH-E message link
     rmail:folder                              RMAIL folder link
     rmail:folder#id                           RMAIL message link
     gnus:group                                Gnus group link
     gnus:group#id                             Gnus article link
     bbdb:R.*Stallman                          BBDB link (with regexp)
     irc:/                   IRC link
     info:org#External links                   Info node link
     shell:ls *.org                            A shell command
     elisp:org-agenda                          Interactive Elisp command
     elisp:(find-file-other-frame "") Elisp form to evaluate

On top of these built-in link types, some are available through the contrib/ directory (see Installation). For example, these links to VM or Wanderlust messages are available when you load the corresponding libraries from the contrib/ directory:

     vm:folder                                 VM folder link
     vm:folder#id                              VM message link
     vm://      VM on remote machine
     vm-imap:account:folder                    VM IMAP folder link
     vm-imap:account:folder#id                 VM IMAP message link
     wl:folder                                 WANDERLUST folder link
     wl:folder#id                              WANDERLUST message link

For customizing Org to add new link types Adding hyperlink types.

A link should be enclosed in double brackets and may contain a descriptive text to be displayed instead of the URL (see Link format), for example:

     [[][GNU Emacs]]

If the description is a file name or URL that points to an image, HTML export (see HTML export) will inline the image as a clickable button. If there is no description at all and the link points to an image, that image will be inlined into the exported HTML file.

Org also finds external links in the normal text and activates them as links. If spaces must be part of the link (for example in ‘bbdb:Richard Stallman’), or if you need to remove ambiguities about the end of the link, enclose them in square brackets.

Next: , Previous: External links, Up: Hyperlinks

4.4 Handling links

Org provides methods to create a link in the correct syntax, to insert it into an Org file, and to follow the link.

C-c l     (org-store-link)
Store a link to the current location. This is a global command (you must create the key binding yourself) which can be used in any buffer to create a link. The link will be stored for later insertion into an Org buffer (see below). What kind of link will be created depends on the current buffer:

Org mode buffers
For Org files, if there is a ‘<<target>>’ at the cursor, the link points to the target. Otherwise it points to the current headline, which will also be the description38.

If the headline has a CUSTOM_ID property, a link to this custom ID will be stored. In addition or alternatively (depending on the value of org-id-link-to-org-use-id), a globally unique ID property will be created and/or used to construct a link39. So using this command in Org buffers will potentially create two links: a human-readable from the custom ID, and one that is globally unique and works even if the entry is moved from file to file. Later, when inserting the link, you need to decide which one to use.

Email/News clients: VM, Rmail, Wanderlust, MH-E, Gnus
Pretty much all Emacs mail clients are supported. The link will point to the current article, or, in some GNUS buffers, to the group. The description is constructed from the author and the subject.

Web browsers: W3 and W3M
Here the link will be the current URL, with the page title as description.

Contacts: BBDB
Links created in a BBDB buffer will point to the current entry.

Chat: IRC
For IRC links, if you set the option org-irc-link-to-logs to t, a ‘file:/’ style link to the relevant point in the logs for the current conversation is created. Otherwise an ‘irc:/’ style link to the user/channel/server under the point will be stored.

Other files
For any other files, the link will point to the file, with a search string (see Search options) pointing to the contents of the current line. If there is an active region, the selected words will form the basis of the search string. If the automatically created link is not working correctly or accurately enough, you can write custom functions to select the search string and to do the search for particular file types—see Custom searches. The key binding C-c l is only a suggestion—see Installation.

Agenda view
When the cursor is in an agenda view, the created link points to the entry referenced by the current line.

C-c C-l     (org-insert-link)
Insert a link40. This prompts for a link to be inserted into the buffer. You can just type a link, using text for an internal link, or one of the link type prefixes mentioned in the examples above. The link will be inserted into the buffer41, along with a descriptive text. If some text was selected when this command is called, the selected text becomes the default description.

Inserting stored links
All links stored during the current session are part of the history for this prompt, so you can access them with <up> and <down> (or M-p/n).

Completion support
Completion with <TAB> will help you to insert valid link prefixes like ‘http:’ or ‘ftp:’, including the prefixes defined through link abbreviations (see Link abbreviations). If you press <RET> after inserting only the prefix, Org will offer specific completion support for some link types42 For example, if you type file <RET>, file name completion (alternative access: C-u C-c C-l, see below) will be offered, and after bbdb <RET> you can complete contact names.

C-u C-c C-l
When C-c C-l is called with a C-u prefix argument, a link to a file will be inserted and you may use file name completion to select the name of the file. The path to the file is inserted relative to the directory of the current Org file, if the linked file is in the current directory or in a sub-directory of it, or if the path is written relative to the current directory using ‘../’. Otherwise an absolute path is used, if possible with ‘~/’ for your home directory. You can force an absolute path with two C-u prefixes.
C-c C-l (with cursor on existing link)
When the cursor is on an existing link, C-c C-l allows you to edit the link and description parts of the link.
C-c C-o     (org-open-at-point)
Open link at point. This will launch a web browser for URLs (using browse-url-at-point), run VM/MH-E/Wanderlust/Rmail/Gnus/BBDB for the corresponding links, and execute the command in a shell link. When the cursor is on an internal link, this command runs the corresponding search. When the cursor is on a TAG list in a headline, it creates the corresponding TAGS view. If the cursor is on a timestamp, it compiles the agenda for that date. Furthermore, it will visit text and remote files in ‘file:’ links with Emacs and select a suitable application for local non-text files. Classification of files is based on file extension only. See option org-file-apps. If you want to override the default application and visit the file with Emacs, use a C-u prefix. If you want to avoid opening in Emacs, use a C-u C-u prefix.
If the cursor is on a headline, but not on a link, offer all links in the headline and entry text. If you want to setup the frame configuration for following links, customize org-link-frame-setup.

When org-return-follows-link is set, <RET> will also follow the link at point.
On links, mouse-2 will open the link just as C-c C-o would. Under Emacs 22 and later, mouse-1 will also follow a link.
Like mouse-2, but force file links to be opened with Emacs, and internal links to be displayed in another window43.
C-c C-x C-v     (org-toggle-inline-images)
Toggle the inline display of linked images. Normally this will only inline images that have no description part in the link, i.e., images that will also be inlined during export. When called with a prefix argument, also display images that do have a link description. You can ask for inline images to be displayed at startup by configuring the variable org-startup-with-inline-images44.
C-c %     (org-mark-ring-push)
Push the current position onto the mark ring, to be able to return easily. Commands following an internal link do this automatically.
C-c &     (org-mark-ring-goto)
Jump back to a recorded position. A position is recorded by the commands following internal links, and by C-c %. Using this command several times in direct succession moves through a ring of previously recorded positions.
C-c C-x C-n     (org-next-link)
C-c C-x C-p     (org-previous-link)
Move forward/backward to the next link in the buffer. At the limit of the buffer, the search fails once, and then wraps around. The key bindings for this are really too long; you might want to bind this also to C-n and C-p
          (add-hook 'org-load-hook
            (lambda ()
              (define-key org-mode-map "\C-n" 'org-next-link)
              (define-key org-mode-map "\C-p" 'org-previous-link)))

Next: , Previous: Handling links, Up: Hyperlinks

4.5 Using links outside Org

You can insert and follow links that have Org syntax not only in Org, but in any Emacs buffer. For this, you should create two global commands, like this (please select suitable global keys yourself):

     (global-set-key "\C-c L" 'org-insert-link-global)
     (global-set-key "\C-c o" 'org-open-at-point-global)

Next: , Previous: Using links outside Org, Up: Hyperlinks

4.6 Link abbreviations

Long URLs can be cumbersome to type, and often many similar links are needed in a document. For this you can use link abbreviations. An abbreviated link looks like this


where the tag is optional. The linkword must be a word, starting with a letter, followed by letters, numbers, ‘-’, and ‘_’. Abbreviations are resolved according to the information in the variable org-link-abbrev-alist that relates the linkwords to replacement text. Here is an example:

     (setq org-link-abbrev-alist
       '(("bugzilla"  . "")
         ("url-to-ja" . "")
         ("google"    . "")
         ("gmap"      . "")
         ("omap"      . "")
         ("ads"       . "")))

If the replacement text contains the string ‘%s’, it will be replaced with the tag. Using ‘%h’ instead of ‘%s’ will url-encode the tag (see the example above, where we need to encode the URL parameter.) Using ‘%(my-function)’ will pass the tag to a custom function, and replace it by the resulting string.

If the replacement text don't contain any specifier, it will simply be appended to the string in order to create the link.

Instead of a string, you may also specify a function that will be called with the tag as the only argument to create the link.

With the above setting, you could link to a specific bug with [[bugzilla:129]], search the web for ‘OrgMode’ with [[google:OrgMode]], show the map location of the Free Software Foundation [[gmap:51 Franklin Street, Boston]] or of Carsten office [[omap:Science Park 904, Amsterdam, The Netherlands]] and find out what the Org author is doing besides Emacs hacking with [[ads:Dominik,C]].

If you need special abbreviations just for a single Org buffer, you can define them in the file with

     #+LINK: bugzilla
     #+LINK: google

In-buffer completion (see Completion) can be used after ‘[’ to complete link abbreviations. You may also define a function org-PREFIX-complete-link that implements special (e.g., completion) support for inserting such a link with C-c C-l. Such a function should not accept any arguments, and return the full link with prefix.

Next: , Previous: Link abbreviations, Up: Hyperlinks

4.7 Search options in file links

File links can contain additional information to make Emacs jump to a particular location in the file when following a link. This can be a line number or a search option after a double45 colon. For example, when the command C-c l creates a link (see Handling links) to a file, it encodes the words in the current line as a search string that can be used to find this line back later when following the link with C-c C-o.

Here is the syntax of the different ways to attach a search to a file link, together with an explanation:

     [[file:~/ Target]]
     [[file:~/*My Target]]
Jump to line 255.
My Target
Search for a link target ‘<<My Target>>’, or do a text search for ‘my target’, similar to the search in internal links, see Internal links. In HTML export (see HTML export), such a file link will become an HTML reference to the corresponding named anchor in the linked file.
*My Target
In an Org file, restrict search to headlines.
Link to a heading with a CUSTOM_ID property
Do a regular expression search for regexp. This uses the Emacs command occur to list all matches in a separate window. If the target file is in Org mode, org-occur is used to create a sparse tree with the matches.

As a degenerate case, a file link with an empty file name can be used to search the current file. For example, [[file:::find me]] does a search for ‘find me’ in the current file, just as ‘[[find me]]’ would.

Previous: Search options, Up: Hyperlinks

4.8 Custom Searches

The default mechanism for creating search strings and for doing the actual search related to a file link may not work correctly in all cases. For example, BibTeX database files have many entries like ‘year="1993"’ which would not result in good search strings, because the only unique identification for a BibTeX entry is the citation key.

If you come across such a problem, you can write custom functions to set the right search string for a particular file type, and to do the search for the string in the file. Using add-hook, these functions need to be added to the hook variables org-create-file-search-functions and org-execute-file-search-functions. See the docstring for these variables for more information. Org actually uses this mechanism for BibTeX database files, and you can use the corresponding code as an implementation example. See the file org-bibtex.el.

Next: , Previous: Hyperlinks, Up: Top

5 TODO items

Org mode does not maintain TODO lists as separate documents46. Instead, TODO items are an integral part of the notes file, because TODO items usually come up while taking notes! With Org mode, simply mark any entry in a tree as being a TODO item. In this way, information is not duplicated, and the entire context from which the TODO item emerged is always present.

Of course, this technique for managing TODO items scatters them throughout your notes file. Org mode compensates for this by providing methods to give you an overview of all the things that you have to do.

Next: , Previous: TODO Items, Up: TODO Items

5.1 Basic TODO functionality

Any headline becomes a TODO item when it starts with the word ‘TODO’, for example:

     *** TODO Write letter to Sam Fortune

The most important commands to work with TODO entries are:

C-c C-t     (org-todo)
Rotate the TODO state of the current item among
          ,-> (unmarked) -> TODO -> DONE --.

If TODO keywords have fast access keys (see Fast access to TODO states), you will be prompted for a TODO keyword through the fast selection interface; this is the default behavior when org-use-fast-todo-selection is non-nil.

The same rotation can also be done “remotely” from the timeline and agenda buffers with the t command key (see Agenda commands).

C-u C-c C-t
When TODO keywords have no selection keys, select a specific keyword using completion; otherwise force cycling through TODO states with no prompt. When org-use-fast-todo-selection is set to prefix, use the fast selection interface.

S-<right> / S-<left>
Select the following/preceding TODO state, similar to cycling. Useful mostly if more than two TODO states are possible (see TODO extensions). See also Conflicts, for a discussion of the interaction with shift-selection-mode. See also the variable org-treat-S-cursor-todo-selection-as-state-change.
C-c / t     (org-show-todo-tree)
View TODO items in a sparse tree (see Sparse trees). Folds the entire buffer, but shows all TODO items (with not-DONE state) and the headings hierarchy above them. With a prefix argument (or by using C-c / T), search for a specific TODO. You will be prompted for the keyword, and you can also give a list of keywords like KWD1|KWD2|... to list entries that match any one of these keywords. With a numeric prefix argument N, show the tree for the Nth keyword in the option org-todo-keywords. With two prefix arguments, find all TODO states, both un-done and done.
C-c a t     (org-todo-list)
Show the global TODO list. Collects the TODO items (with not-DONE states) from all agenda files (see Agenda Views) into a single buffer. The new buffer will be in agenda-mode, which provides commands to examine and manipulate the TODO entries from the new buffer (see Agenda commands). See Global TODO list, for more information.
S-M-<RET>     (org-insert-todo-heading)
Insert a new TODO entry below the current one.

Changing a TODO state can also trigger tag changes. See the docstring of the option org-todo-state-tags-triggers for details.

Next: , Previous: TODO basics, Up: TODO Items

5.2 Extended use of TODO keywords

By default, marked TODO entries have one of only two states: TODO and DONE. Org mode allows you to classify TODO items in more complex ways with TODO keywords (stored in org-todo-keywords). With special setup, the TODO keyword system can work differently in different files.

Note that tags are another way to classify headlines in general and TODO items in particular (see Tags).

Next: , Previous: TODO extensions, Up: TODO extensions

5.2.1 TODO keywords as workflow states

You can use TODO keywords to indicate different sequential states in the process of working on an item, for example47:

     (setq org-todo-keywords
       '((sequence "TODO" "FEEDBACK" "VERIFY" "|" "DONE" "DELEGATED")))

The vertical bar separates the TODO keywords (states that need action) from the DONE states (which need no further action). If you don't provide the separator bar, the last state is used as the DONE state. With this setup, the command C-c C-t will cycle an entry from TODO to FEEDBACK, then to VERIFY, and finally to DONE and DELEGATED. You may also use a numeric prefix argument to quickly select a specific state. For example C-3 C-c C-t will change the state immediately to VERIFY. Or you can use S-<left> to go backward through the sequence. If you define many keywords, you can use in-buffer completion (see Completion) or even a special one-key selection scheme (see Fast access to TODO states) to insert these words into the buffer. Changing a TODO state can be logged with a timestamp, see Tracking TODO state changes, for more information.

Next: , Previous: Workflow states, Up: TODO extensions

5.2.2 TODO keywords as types

The second possibility is to use TODO keywords to indicate different types of action items. For example, you might want to indicate that items are for “work” or “home”. Or, when you work with several people on a single project, you might want to assign action items directly to persons, by using their names as TODO keywords. This would be set up like this:

     (setq org-todo-keywords '((type "Fred" "Sara" "Lucy" "|" "DONE")))

In this case, different keywords do not indicate a sequence, but rather different types. So the normal work flow would be to assign a task to a person, and later to mark it DONE. Org mode supports this style by adapting the workings of the command C-c C-t48. When used several times in succession, it will still cycle through all names, in order to first select the right type for a task. But when you return to the item after some time and execute C-c C-t again, it will switch from any name directly to DONE. Use prefix arguments or completion to quickly select a specific name. You can also review the items of a specific TODO type in a sparse tree by using a numeric prefix to C-c / t. For example, to see all things Lucy has to do, you would use C-3 C-c / t. To collect Lucy's items from all agenda files into a single buffer, you would use the numeric prefix argument as well when creating the global TODO list: C-3 C-c a t.

Next: , Previous: TODO types, Up: TODO extensions

5.2.3 Multiple keyword sets in one file

Sometimes you may want to use different sets of TODO keywords in parallel. For example, you may want to have the basic TODO/DONE, but also a workflow for bug fixing, and a separate state indicating that an item has been canceled (so it is not DONE, but also does not require action). Your setup would then look like this:

     (setq org-todo-keywords
           '((sequence "TODO" "|" "DONE")
             (sequence "REPORT" "BUG" "KNOWNCAUSE" "|" "FIXED")
             (sequence "|" "CANCELED")))

The keywords should all be different, this helps Org mode to keep track of which subsequence should be used for a given entry. In this setup, C-c C-t only operates within a subsequence, so it switches from DONE to (nothing) to TODO, and from FIXED to (nothing) to REPORT. Therefore you need a mechanism to initially select the correct sequence. Besides the obvious ways like typing a keyword or using completion, you may also apply the following commands:

C-u C-u C-c C-t
These keys jump from one TODO subset to the next. In the above example, C-u C-u C-c C-t or C-S-<right> would jump from TODO or DONE to REPORT, and any of the words in the second row to CANCELED. Note that the C-S- key binding conflict with shift-selection-mode (see Conflicts).
S-<<left>> and S-<<right>> and walk through all keywords from all sets, so for example S-<<right>> would switch from DONE to REPORT in the example above. See also Conflicts, for a discussion of the interaction with shift-selection-mode.

Next: , Previous: Multiple sets in one file, Up: TODO extensions

5.2.4 Fast access to TODO states

If you would like to quickly change an entry to an arbitrary TODO state instead of cycling through the states, you can set up keys for single-letter access to the states. This is done by adding the selection character after each keyword, in parentheses49. For example:

     (setq org-todo-keywords
           '((sequence "TODO(t)" "|" "DONE(d)")
             (sequence "REPORT(r)" "BUG(b)" "KNOWNCAUSE(k)" "|" "FIXED(f)")
             (sequence "|" "CANCELED(c)")))

If you then press C-c C-t followed by the selection key, the entry will be switched to this state. SPC can be used to remove any TODO keyword from an entry.50

Next: , Previous: Fast access to TODO states, Up: TODO extensions

5.2.5 Setting up keywords for individual files

It can be very useful to use different aspects of the TODO mechanism in different files. For file-local settings, you need to add special lines to the file which set the keywords and interpretation for that file only. For example, to set one of the two examples discussed above, you need one of the following lines, starting in column zero anywhere in the file:


(you may also write #+SEQ_TODO to be explicit about the interpretation, but it means the same as #+TODO), or

     #+TYP_TODO: Fred Sara Lucy Mike | DONE

A setup for using several sets in parallel would be:

     #+TODO: TODO | DONE

To make sure you are using the correct keyword, type ‘#+’ into the buffer and then use M-<TAB> completion.

Remember that the keywords after the vertical bar (or the last keyword if no bar is there) must always mean that the item is DONE (although you may use a different word). After changing one of these lines, use C-c C-c with the cursor still in the line to make the changes known to Org mode51.

Next: , Previous: Per-file keywords, Up: TODO extensions

5.2.6 Faces for TODO keywords

Org mode highlights TODO keywords with special faces: org-todo for keywords indicating that an item still has to be acted upon, and org-done for keywords indicating that an item is finished. If you are using more than 2 different states, you might want to use special faces for some of them. This can be done using the option org-todo-keyword-faces. For example:

     (setq org-todo-keyword-faces
           '(("TODO" . org-warning) ("STARTED" . "yellow")
             ("CANCELED" . (:foreground "blue" :weight bold))))

While using a list with face properties as shown for CANCELED should work, this does not always seem to be the case. If necessary, define a special face and use that. A string is interpreted as a color. The option org-faces-easy-properties determines if that color is interpreted as a foreground or a background color.

Previous: Faces for TODO keywords, Up: TODO extensions

5.2.7 TODO dependencies

The structure of Org files (hierarchy and lists) makes it easy to define TODO dependencies. Usually, a parent TODO task should not be marked DONE until all subtasks (defined as children tasks) are marked as DONE. And sometimes there is a logical sequence to a number of (sub)tasks, so that one task cannot be acted upon before all siblings above it are done. If you customize the option org-enforce-todo-dependencies, Org will block entries from changing state to DONE while they have children that are not DONE. Furthermore, if an entry has a property ORDERED, each of its children will be blocked until all earlier siblings are marked DONE. Here is an example:

     * TODO Blocked until (two) is done
     ** DONE one
     ** TODO two
     * Parent
       :ORDERED: t
     ** TODO a
     ** TODO b, needs to wait for (a)
     ** TODO c, needs to wait for (a) and (b)
C-c C-x o     (org-toggle-ordered-property)
Toggle the ORDERED property of the current entry. A property is used for this behavior because this should be local to the current entry, not inherited like a tag. However, if you would like to track the value of this property with a tag for better visibility, customize the option org-track-ordered-property-with-tag.
C-u C-u C-u C-c C-t
Change TODO state, circumventing any state blocking.

If you set the option org-agenda-dim-blocked-tasks, TODO entries that cannot be closed because of such dependencies will be shown in a dimmed font or even made invisible in agenda views (see Agenda Views).

You can also block changes of TODO states by looking at checkboxes (see Checkboxes). If you set the option org-enforce-todo-checkbox-dependencies, an entry that has unchecked checkboxes will be blocked from switching to DONE.

If you need more complex dependency structures, for example dependencies between entries in different trees or files, check out the contributed module org-depend.el.

Next: , Previous: TODO extensions, Up: TODO Items

5.3 Progress logging

Org mode can automatically record a timestamp and possibly a note when you mark a TODO item as DONE, or even each time you change the state of a TODO item. This system is highly configurable; settings can be on a per-keyword basis and can be localized to a file or even a subtree. For information on how to clock working time for a task, see Clocking work time.

Next: , Previous: Progress logging, Up: Progress logging

5.3.1 Closing items

The most basic logging is to keep track of when a certain TODO item was finished. This is achieved with52

     (setq org-log-done 'time)

Then each time you turn an entry from a TODO (not-done) state into any of the DONE states, a line ‘CLOSED: [timestamp]’ will be inserted just after the headline. If you turn the entry back into a TODO item through further state cycling, that line will be removed again. If you turn the entry back to a non-TODO state (by pressing <C-c C-t SPC> for example), that line will also be removed, unless you set org-closed-keep-when-no-todo to non-nil. If you want to record a note along with the timestamp, use53

     (setq org-log-done 'note)

You will then be prompted for a note, and that note will be stored below the entry with a ‘Closing Note’ heading.

In the timeline (see Timeline) and in the agenda (see Weekly/daily agenda), you can then use the l key to display the TODO items with a ‘CLOSED’ timestamp on each day, giving you an overview of what has been done.

Next: , Previous: Closing items, Up: Progress logging

5.3.2 Tracking TODO state changes

When TODO keywords are used as workflow states (see Workflow states), you might want to keep track of when a state change occurred and maybe take a note about this change. You can either record just a timestamp, or a time-stamped note for a change. These records will be inserted after the headline as an itemized list, newest first54. When taking a lot of notes, you might want to get the notes out of the way into a drawer (see Drawers). Customize org-log-into-drawer to get this behavior—the recommended drawer for this is called LOGBOOK55. You can also overrule the setting of this variable for a subtree by setting a LOG_INTO_DRAWER property.

Since it is normally too much to record a note for every state, Org mode expects configuration on a per-keyword basis for this. This is achieved by adding special markers ‘!’ (for a timestamp) or ‘@’ (for a note with timestamp) in parentheses after each keyword. For example, with the setting

     (setq org-todo-keywords
       '((sequence "TODO(t)" "WAIT(w@/!)" "|" "DONE(d!)" "CANCELED(c@)")))

To record a timestamp without a note for TODO keywords configured with ‘@’, just type C-c C-c to enter a blank note when prompted.

You not only define global TODO keywords and fast access keys, but also request that a time is recorded when the entry is set to DONE56, and that a note is recorded when switching to WAIT or CANCELED. The setting for WAIT is even more special: the ‘!’ after the slash means that in addition to the note taken when entering the state, a timestamp should be recorded when leaving the WAIT state, if and only if the target state does not configure logging for entering it. So it has no effect when switching from WAIT to DONE, because DONE is configured to record a timestamp only. But when switching from WAIT back to TODO, the ‘/!’ in the WAIT setting now triggers a timestamp even though TODO has no logging configured.

You can use the exact same syntax for setting logging preferences local to a buffer:

     #+TODO: TODO(t) WAIT(w@/!) | DONE(d!) CANCELED(c@)

In order to define logging settings that are local to a subtree or a single item, define a LOGGING property in this entry. Any non-empty LOGGING property resets all logging settings to nil. You may then turn on logging for this specific tree using STARTUP keywords like lognotedone or logrepeat, as well as adding state specific settings like TODO(!). For example

     * TODO Log each state with only a time
     * TODO Only log when switching to WAIT, and when repeating
       :LOGGING: WAIT(@) logrepeat
     * TODO No logging at all
       :LOGGING: nil

Previous: Tracking TODO state changes, Up: Progress logging

5.3.3 Tracking your habits

Org has the ability to track the consistency of a special category of TODOs, called “habits”. A habit has the following properties:

  1. You have enabled the habits module by customizing org-modules.
  2. The habit is a TODO item, with a TODO keyword representing an open state.
  3. The property STYLE is set to the value habit.
  4. The TODO has a scheduled date, usually with a .+ style repeat interval. A ++ style may be appropriate for habits with time constraints, e.g., must be done on weekends, or a + style for an unusual habit that can have a backlog, e.g., weekly reports.
  5. The TODO may also have minimum and maximum ranges specified by using the syntax ‘.+2d/3d’, which says that you want to do the task at least every three days, but at most every two days.
  6. You must also have state logging for the DONE state enabled (see Tracking TODO state changes), in order for historical data to be represented in the consistency graph. If it is not enabled it is not an error, but the consistency graphs will be largely meaningless.

To give you an idea of what the above rules look like in action, here's an actual habit with some history:

     ** TODO Shave
        SCHEDULED: <2009-10-17 Sat .+2d/4d>
        - State "DONE"       from "TODO"       [2009-10-15 Thu]
        - State "DONE"       from "TODO"       [2009-10-12 Mon]
        - State "DONE"       from "TODO"       [2009-10-10 Sat]
        - State "DONE"       from "TODO"       [2009-10-04 Sun]
        - State "DONE"       from "TODO"       [2009-10-02 Fri]
        - State "DONE"       from "TODO"       [2009-09-29 Tue]
        - State "DONE"       from "TODO"       [2009-09-25 Fri]
        - State "DONE"       from "TODO"       [2009-09-19 Sat]
        - State "DONE"       from "TODO"       [2009-09-16 Wed]
        - State "DONE"       from "TODO"       [2009-09-12 Sat]
        :STYLE:    habit
        :LAST_REPEAT: [2009-10-19 Mon 00:36]

What this habit says is: I want to shave at most every 2 days (given by the SCHEDULED date and repeat interval) and at least every 4 days. If today is the 15th, then the habit first appears in the agenda on Oct 17, after the minimum of 2 days has elapsed, and will appear overdue on Oct 19, after four days have elapsed.

What's really useful about habits is that they are displayed along with a consistency graph, to show how consistent you've been at getting that task done in the past. This graph shows every day that the task was done over the past three weeks, with colors for each day. The colors used are:

If the task wasn't to be done yet on that day.
If the task could have been done on that day.
If the task was going to be overdue the next day.
If the task was overdue on that day.

In addition to coloring each day, the day is also marked with an asterisk if the task was actually done that day, and an exclamation mark to show where the current day falls in the graph.

There are several configuration variables that can be used to change the way habits are displayed in the agenda.

The buffer column at which the consistency graph should be drawn. This will overwrite any text in that column, so it is a good idea to keep your habits' titles brief and to the point.
The amount of history, in days before today, to appear in consistency graphs.
The number of days after today that will appear in consistency graphs.
If non-nil, only show habits in today's agenda view. This is set to true by default.

Lastly, pressing K in the agenda buffer will cause habits to temporarily be disabled and they won't appear at all. Press K again to bring them back. They are also subject to tag filtering, if you have habits which should only be done in certain contexts, for example.

Next: , Previous: Progress logging, Up: TODO Items

5.4 Priorities

If you use Org mode extensively, you may end up with enough TODO items that it starts to make sense to prioritize them. Prioritizing can be done by placing a priority cookie into the headline of a TODO item, like this

     *** TODO [#A] Write letter to Sam Fortune

By default, Org mode supports three priorities: ‘A’, ‘B’, and ‘C’. ‘A’ is the highest priority. An entry without a cookie is treated just like priority ‘B’. Priorities make a difference only for sorting in the agenda (see Weekly/daily agenda); outside the agenda, they have no inherent meaning to Org mode. The cookies can be highlighted with special faces by customizing org-priority-faces.

Priorities can be attached to any outline node; they do not need to be TODO items.

C-c ,
Set the priority of the current headline (org-priority). The command prompts for a priority character ‘A’, ‘B’ or ‘C’. When you press <SPC> instead, the priority cookie is removed from the headline. The priorities can also be changed “remotely” from the timeline and agenda buffer with the , command (see Agenda commands).
S-<up>     (org-priority-up)
S-<down>     (org-priority-down)
Increase/decrease priority of current headline57. Note that these keys are also used to modify timestamps (see Creating timestamps). See also Conflicts, for a discussion of the interaction with shift-selection-mode.

You can change the range of allowed priorities by setting the options org-highest-priority, org-lowest-priority, and org-default-priority. For an individual buffer, you may set these values (highest, lowest, default) like this (please make sure that the highest priority is earlier in the alphabet than the lowest priority):


Next: , Previous: Priorities, Up: TODO Items

5.5 Breaking tasks down into subtasks

It is often advisable to break down large tasks into smaller, manageable subtasks. You can do this by creating an outline tree below a TODO item, with detailed subtasks on the tree58. To keep the overview over the fraction of subtasks that are already completed, insert either ‘[/]’ or ‘[%]’ anywhere in the headline. These cookies will be updated each time the TODO status of a child changes, or when pressing C-c C-c on the cookie. For example:

     * Organize Party [33%]
     ** TODO Call people [1/2]
     *** TODO Peter
     *** DONE Sarah
     ** TODO Buy food
     ** DONE Talk to neighbor

If a heading has both checkboxes and TODO children below it, the meaning of the statistics cookie become ambiguous. Set the property COOKIE_DATA to either ‘checkbox’ or ‘todo’ to resolve this issue.

If you would like to have the statistics cookie count any TODO entries in the subtree (not just direct children), configure org-hierarchical-todo-statistics. To do this for a single subtree, include the word ‘recursive’ into the value of the COOKIE_DATA property.

     * Parent capturing statistics [2/20]
       :COOKIE_DATA: todo recursive

If you would like a TODO entry to automatically change to DONE when all children are done, you can use the following setup:

     (defun org-summary-todo (n-done n-not-done)
       "Switch entry to DONE when all subentries are done, to TODO otherwise."
       (let (org-log-done org-log-states)   ; turn off logging
         (org-todo (if (= n-not-done 0) "DONE" "TODO"))))
     (add-hook 'org-after-todo-statistics-hook 'org-summary-todo)

Another possibility is the use of checkboxes to identify (a hierarchy of) a large number of subtasks (see Checkboxes).

Previous: Breaking down tasks, Up: TODO Items

5.6 Checkboxes

Every item in a plain list59 (see Plain lists) can be made into a checkbox by starting it with the string ‘[ ]’. This feature is similar to TODO items (see TODO Items), but is more lightweight. Checkboxes are not included in the global TODO list, so they are often great to split a task into a number of simple steps. Or you can use them in a shopping list. To toggle a checkbox, use C-c C-c, or use the mouse (thanks to Piotr Zielinski's org-mouse.el).

Here is an example of a checkbox list.

     * TODO Organize party [2/4]
       - [-] call people [1/3]
         - [ ] Peter
         - [X] Sarah
         - [ ] Sam
       - [X] order food
       - [ ] think about what music to play
       - [X] talk to the neighbors

Checkboxes work hierarchically, so if a checkbox item has children that are checkboxes, toggling one of the children checkboxes will make the parent checkbox reflect if none, some, or all of the children are checked.

The ‘[2/4]’ and ‘[1/3]’ in the first and second line are cookies indicating how many checkboxes present in this entry have been checked off, and the total number of checkboxes present. This can give you an idea on how many checkboxes remain, even without opening a folded entry. The cookies can be placed into a headline or into (the first line of) a plain list item. Each cookie covers checkboxes of direct children structurally below the headline/item on which the cookie appears60. You have to insert the cookie yourself by typing either ‘[/]’ or ‘[%]’. With ‘[/]’ you get an ‘n out of m’ result, as in the examples above. With ‘[%]’ you get information about the percentage of checkboxes checked (in the above example, this would be ‘[50%]’ and ‘[33%]’, respectively). In a headline, a cookie can count either checkboxes below the heading or TODO states of children, and it will display whatever was changed last. Set the property COOKIE_DATA to either ‘checkbox’ or ‘todo’ to resolve this issue.

If the current outline node has an ORDERED property, checkboxes must be checked off in sequence, and an error will be thrown if you try to check off a box while there are unchecked boxes above it.

The following commands work with checkboxes:

C-c C-c     (org-toggle-checkbox)
Toggle checkbox status or (with prefix arg) checkbox presence at point. With a single prefix argument, add an empty checkbox or remove the current one61. With a double prefix argument, set it to ‘[-]’, which is considered to be an intermediate state.
C-c C-x C-b     (org-toggle-checkbox)
Toggle checkbox status or (with prefix arg) checkbox presence at point. With double prefix argument, set it to ‘[-]’, which is considered to be an intermediate state.
  • If there is an active region, toggle the first checkbox in the region and set all remaining boxes to the same status as the first. With a prefix arg, add or remove the checkbox for all items in the region.
  • If the cursor is in a headline, toggle checkboxes in the region between this headline and the next (so not the entire subtree).
  • If there is no active region, just toggle the checkbox at point.

M-S-<RET>     (org-insert-todo-heading)
Insert a new item with a checkbox. This works only if the cursor is already in a plain list item (see Plain lists).
C-c C-x o     (org-toggle-ordered-property)
Toggle the ORDERED property of the entry, to toggle if checkboxes must be checked off in sequence. A property is used for this behavior because this should be local to the current entry, not inherited like a tag. However, if you would like to track the value of this property with a tag for better visibility, customize org-track-ordered-property-with-tag.
C-c #     (org-update-statistics-cookies)
Update the statistics cookie in the current outline entry. When called with a C-u prefix, update the entire file. Checkbox statistic cookies are updated automatically if you toggle checkboxes with C-c C-c and make new ones with M-S-<RET>. TODO statistics cookies update when changing TODO states. If you delete boxes/entries or add/change them by hand, use this command to get things back into sync.

Next: , Previous: TODO Items, Up: Top

6 Tags

An excellent way to implement labels and contexts for cross-correlating information is to assign tags to headlines. Org mode has extensive support for tags.

Every headline can contain a list of tags; they occur at the end of the headline. Tags are normal words containing letters, numbers, ‘_’, and ‘@’. Tags must be preceded and followed by a single colon, e.g., ‘:work:’. Several tags can be specified, as in ‘:work:urgent:’. Tags will by default be in bold face with the same color as the headline. You may specify special faces for specific tags using the option org-tag-faces, in much the same way as you can for TODO keywords (see Faces for TODO keywords).

Next: , Previous: Tags, Up: Tags

6.1 Tag inheritance

Tags make use of the hierarchical structure of outline trees. If a heading has a certain tag, all subheadings will inherit the tag as well. For example, in the list

     * Meeting with the French group      :work:
     ** Summary by Frank                  :boss:notes:
     *** TODO Prepare slides for him      :action:

the final heading will have the tags ‘:work:’, ‘:boss:’, ‘:notes:’, and ‘:action:’ even though the final heading is not explicitly marked with those tags. You can also set tags that all entries in a file should inherit just as if these tags were defined in a hypothetical level zero that surrounds the entire file. Use a line like this62:

     #+FILETAGS: :Peter:Boss:Secret:

To limit tag inheritance to specific tags, use org-tags-exclude-from-inheritance. To turn it off entirely, use org-use-tag-inheritance.

When a headline matches during a tags search while tag inheritance is turned on, all the sublevels in the same tree will (for a simple match form) match as well63. The list of matches may then become very long. If you only want to see the first tags match in a subtree, configure org-tags-match-list-sublevels (not recommended).

Tag inheritance is relevant when the agenda search tries to match a tag, either in the tags or tags-todo agenda types. In other agenda types, org-use-tag-inheritance has no effect. Still, you may want to have your tags correctly set in the agenda, so that tag filtering works fine, with inherited tags. Set org-agenda-use-tag-inheritance to control this: the default value includes all agenda types, but setting this to nil can really speed up agenda generation.

Next: , Previous: Tag inheritance, Up: Tags

6.2 Setting tags

Tags can simply be typed into the buffer at the end of a headline. After a colon, M-<TAB> offers completion on tags. There is also a special command for inserting tags:

C-c C-q     (org-set-tags-command)
Enter new tags for the current headline. Org mode will either offer completion or a special single-key interface for setting tags, see below. After pressing <RET>, the tags will be inserted and aligned to org-tags-column. When called with a C-u prefix, all tags in the current buffer will be aligned to that column, just to make things look nice. TAGS are automatically realigned after promotion, demotion, and TODO state changes (see TODO basics).

C-c C-c     (org-set-tags-command)
When the cursor is in a headline, this does the same as C-c C-q.

Org supports tag insertion based on a list of tags. By default this list is constructed dynamically, containing all tags currently used in the buffer. You may also globally specify a hard list of tags with the variable org-tag-alist. Finally you can set the default tags for a given file with lines like

     #+TAGS: @work @home @tennisclub
     #+TAGS: laptop car pc sailboat

If you have globally defined your preferred set of tags using the variable org-tag-alist, but would like to use a dynamic tag list in a specific file, add an empty TAGS option line to that file:


If you have a preferred set of tags that you would like to use in every file, in addition to those defined on a per-file basis by TAGS option lines, then you may specify a list of tags with the variable org-tag-persistent-alist. You may turn this off on a per-file basis by adding a STARTUP option line to that file:

     #+STARTUP: noptag

By default Org mode uses the standard minibuffer completion facilities for entering tags. However, it also implements another, quicker, tag selection method called fast tag selection. This allows you to select and deselect tags with just a single key press. For this to work well you should assign unique letters to most of your commonly used tags. You can do this globally by configuring the variable org-tag-alist in your .emacs file. For example, you may find the need to tag many items in different files with ‘:@home:’. In this case you can set something like:

     (setq org-tag-alist '(("@work" . ?w) ("@home" . ?h) ("laptop" . ?l)))

If the tag is only relevant to the file you are working on, then you can instead set the TAGS option line as:

     #+TAGS: @work(w)  @home(h)  @tennisclub(t)  laptop(l)  pc(p)

The tags interface will show the available tags in a splash window. If you want to start a new line after a specific tag, insert ‘\n’ into the tag list

     #+TAGS: @work(w)  @home(h)  @tennisclub(t) \n laptop(l)  pc(p)

or write them in two lines:

     #+TAGS: @work(w)  @home(h)  @tennisclub(t)
     #+TAGS: laptop(l)  pc(p)

You can also group together tags that are mutually exclusive by using braces, as in:

     #+TAGS: { @work(w)  @home(h)  @tennisclub(t) }  laptop(l)  pc(p)

you indicate that at most one of ‘@work’, ‘@home’, and ‘@tennisclub’ should be selected. Multiple such groups are allowed.

Don't forget to press C-c C-c with the cursor in one of these lines to activate any changes.

To set these mutually exclusive groups in the variable org-tag-alist, you must use the dummy tags :startgroup and :endgroup instead of the braces. Similarly, you can use :newline to indicate a line break. The previous example would be set globally by the following configuration:

     (setq org-tag-alist '((:startgroup . nil)
                           ("@work" . ?w) ("@home" . ?h)
                           ("@tennisclub" . ?t)
                           (:endgroup . nil)
                           ("laptop" . ?l) ("pc" . ?p)))

If at least one tag has a selection key then pressing C-c C-c will automatically present you with a special interface, listing inherited tags, the tags of the current headline, and a list of all valid tags with corresponding keys64. In this interface, you can use the following keys:

Pressing keys assigned to tags will add or remove them from the list of tags in the current line. Selecting a tag in a group of mutually exclusive tags will turn off any other tags from that group.
Enter a tag in the minibuffer, even if the tag is not in the predefined list. You will be able to complete on all tags present in the buffer. You can also add several tags: just separate them with a comma.

Clear all tags for this line.
Accept the modified set.
Abort without installing changes.
If q is not assigned to a tag, it aborts like C-g.
Turn off groups of mutually exclusive tags. Use this to (as an exception) assign several tags from such a group.
Toggle auto-exit after the next change (see below). If you are using expert mode, the first C-c will display the selection window.

This method lets you assign tags to a headline with very few keys. With the above setup, you could clear the current tags and set ‘@home’, ‘laptop’ and ‘pc’ tags with just the following keys: C-c C-c <SPC> h l p <RET>. Switching from ‘@home’ to ‘@work’ would be done with C-c C-c w <RET> or alternatively with C-c C-c C-c w. Adding the non-predefined tag ‘Sarah’ could be done with C-c C-c <TAB> S a r a h <RET> <RET>.

If you find that most of the time you need only a single key press to modify your list of tags, set org-fast-tag-selection-single-key. Then you no longer have to press <RET> to exit fast tag selection—it will immediately exit after the first change. If you then occasionally need more keys, press C-c to turn off auto-exit for the current tag selection process (in effect: start selection with C-c C-c C-c instead of C-c C-c). If you set the variable to the value expert, the special window is not even shown for single-key tag selection, it comes up only when you press an extra C-c.

Next: , Previous: Setting tags, Up: Tags

6.3 Tag groups

In a set of mutually exclusive tags, the first tag can be defined as a group tag. When you search for a group tag, it will return matches for all members in the group. In an agenda view, filtering by a group tag will display headlines tagged with at least one of the members of the group. This makes tag searches and filters even more flexible.

You can set group tags by inserting a colon between the group tag and other tags—beware that all whitespaces are mandatory so that Org can parse this line correctly:

     #+TAGS: { @read : @read_book @read_ebook }

In this example, ‘@read’ is a group tag for a set of three tags: ‘@read’, ‘@read_book’ and ‘@read_ebook’.

You can also use the :grouptags keyword directly when setting org-tag-alist:

     (setq org-tag-alist '((:startgroup . nil)
                           ("@read" . nil)
                           (:grouptags . nil)
                           ("@read_book" . nil)
                           ("@read_ebook" . nil)
                           (:endgroup . nil)))

You cannot nest group tags or use a group tag as a tag in another group.

If you want to ignore group tags temporarily, toggle group tags support with org-toggle-tags-groups, bound to C-c C-x q. If you want to disable tag groups completely, set org-group-tags to nil.

Previous: Tag groups, Up: Tags

6.4 Tag searches

Once a system of tags has been set up, it can be used to collect related information into special lists.

C-c / m or C-c \     (org-match-sparse-tree)
Create a sparse tree with all headlines matching a tags/property/TODO search. With a C-u prefix argument, ignore headlines that are not a TODO line. See Matching tags and properties.
C-c a m     (org-tags-view)
Create a global list of tag matches from all agenda files. See Matching tags and properties.
C-c a M     (org-tags-view)
Create a global list of tag matches from all agenda files, but check only TODO items and force checking subitems (see the option org-tags-match-list-sublevels).

These commands all prompt for a match string which allows basic Boolean logic like ‘+boss+urgent-project1’, to find entries with tags ‘boss’ and ‘urgent’, but not ‘project1’, or ‘Kathy|Sally’ to find entries which are tagged, like ‘Kathy’ or ‘Sally’. The full syntax of the search string is rich and allows also matching against TODO keywords, entry levels and properties. For a complete description with many examples, see Matching tags and properties.

Next: , Previous: Tags, Up: Top

7 Properties and columns

A property is a key-value pair associated with an entry. Properties can be set so they are associated with a single entry, with every entry in a tree, or with every entry in an Org mode file.

There are two main applications for properties in Org mode. First, properties are like tags, but with a value. Imagine maintaining a file where you document bugs and plan releases for a piece of software. Instead of using tags like :release_1:, :release_2:, you can use a property, say :Release:, that in different subtrees has different values, such as 1.0 or 2.0. Second, you can use properties to implement (very basic) database capabilities in an Org buffer. Imagine keeping track of your music CDs, where properties could be things such as the album, artist, date of release, number of tracks, and so on.

Properties can be conveniently edited and viewed in column view (see Column view).

Next: , Previous: Properties and Columns, Up: Properties and Columns

7.1 Property syntax

Properties are key-value pairs. When they are associated with a single entry or with a tree they need to be inserted into a special drawer (see Drawers) with the name PROPERTIES. Each property is specified on a single line, with the key (surrounded by colons) first, and the value after it. Here is an example:

     * CD collection
     ** Classic
     *** Goldberg Variations
         :Title:     Goldberg Variations
         :Composer:  J.S. Bach
         :Artist:    Glen Gould
         :Publisher: Deutsche Grammophon
         :NDisks:    1

Depending on the value of org-use-property-inheritance, a property set this way will either be associated with a single entry, or the sub-tree defined by the entry, see Property inheritance.

You may define the allowed values for a particular property ‘:Xyz:’ by setting a property ‘:Xyz_ALL:’. This special property is inherited, so if you set it in a level 1 entry, it will apply to the entire tree. When allowed values are defined, setting the corresponding property becomes easier and is less prone to typing errors. For the example with the CD collection, we can predefine publishers and the number of disks in a box like this:

     * CD collection
       :NDisks_ALL:  1 2 3 4
       :Publisher_ALL: "Deutsche Grammophon" Philips EMI

If you want to set properties that can be inherited by any entry in a file, use a line like

     #+PROPERTY: NDisks_ALL 1 2 3 4

Contrary to properties set from a special drawer, you have to refresh the buffer with C-c C-c to activate this changes.

If you want to add to the value of an existing property, append a + to the property name. The following results in the property var having the value “foo=1 bar=2”.

     #+PROPERTY: var  foo=1
     #+PROPERTY: var+ bar=2

It is also possible to add to the values of inherited properties. The following results in the genres property having the value “Classic Baroque” under the Goldberg Variations subtree.

     * CD collection
     ** Classic
         :GENRES: Classic
     *** Goldberg Variations
         :Title:     Goldberg Variations
         :Composer:  J.S. Bach
         :Artist:    Glen Gould
         :Publisher: Deutsche Grammophon
         :NDisks:    1
         :GENRES+:   Baroque

Note that a property can only have one entry per Drawer.

Property values set with the global variable org-global-properties can be inherited by all entries in all Org files.

The following commands help to work with properties:

M-<TAB>     (pcomplete)
After an initial colon in a line, complete property keys. All keys used in the current file will be offered as possible completions.
C-c C-x p     (org-set-property)
Set a property. This prompts for a property name and a value. If necessary, the property drawer is created as well.
C-u M-x org-insert-drawer RET
Insert a property drawer into the current entry. The drawer will be inserted early in the entry, but after the lines with planning information like deadlines.
C-c C-c     (org-property-action)
With the cursor in a property drawer, this executes property commands.
C-c C-c s     (org-set-property)
Set a property in the current entry. Both the property and the value can be inserted using completion.
S-<right>     (org-property-next-allowed-value)
S-<left>     (org-property-previous-allowed-value)
Switch property at point to the next/previous allowed value.
C-c C-c d     (org-delete-property)
Remove a property from the current entry.
C-c C-c D     (org-delete-property-globally)
Globally remove a property, from all entries in the current file.
C-c C-c c     (org-compute-property-at-point)
Compute the property at point, using the operator and scope from the nearest column format definition.

Next: , Previous: Property syntax, Up: Properties and Columns

7.2 Special properties

Special properties provide an alternative access method to Org mode features, like the TODO state or the priority of an entry, discussed in the previous chapters. This interface exists so that you can include these states in a column view (see Column view), or to use them in queries. The following property names are special and (except for :CATEGORY:) should not be used as keys in the properties drawer:

     ID           A globally unique ID used for synchronization during
                  iCalendar or MobileOrg export.
     TODO         The TODO keyword of the entry.
     TAGS         The tags defined directly in the headline.
     ALLTAGS      All tags, including inherited ones.
     CATEGORY     The category of an entry.
     PRIORITY     The priority of the entry, a string with a single letter.
     DEADLINE     The deadline time string, without the angular brackets.
     SCHEDULED    The scheduling timestamp, without the angular brackets.
     CLOSED       When was this entry closed?
     TIMESTAMP    The first keyword-less timestamp in the entry.
     TIMESTAMP_IA The first inactive timestamp in the entry.
     CLOCKSUM     The sum of CLOCK intervals in the subtree.  org-clock-sum
                  must be run first to compute the values in the current buffer.
     CLOCKSUM_T   The sum of CLOCK intervals in the subtree for today.
                  org-clock-sum-today must be run first to compute the
                  values in the current buffer.
     BLOCKED      "t" if task is currently blocked by children or siblings
     ITEM         The headline of the entry.
     FILE         The filename the entry is located in.

Next: , Previous: Special properties, Up: Properties and Columns

7.3 Property searches

To create sparse trees and special lists with selection based on properties, the same commands are used as for tag searches (see Tag searches).

C-c / m or C-c \     (org-match-sparse-tree)
Create a sparse tree with all matching entries. With a C-u prefix argument, ignore headlines that are not a TODO line.
C-c a m     (org-tags-view)
Create a global list of tag/property matches from all agenda files. See Matching tags and properties.
C-c a M     (org-tags-view)
Create a global list of tag matches from all agenda files, but check only TODO items and force checking of subitems (see the option org-tags-match-list-sublevels).

The syntax for the search string is described in Matching tags and properties.

There is also a special command for creating sparse trees based on a single property:

C-c / p
Create a sparse tree based on the value of a property. This first prompts for the name of a property, and then for a value. A sparse tree is created with all entries that define this property with the given value. If you enclose the value in curly braces, it is interpreted as a regular expression and matched against the property values.

Next: , Previous: Property searches, Up: Properties and Columns

7.4 Property Inheritance

The outline structure of Org mode documents lends itself to an inheritance model of properties: if the parent in a tree has a certain property, the children can inherit this property. Org mode does not turn this on by default, because it can slow down property searches significantly and is often not needed. However, if you find inheritance useful, you can turn it on by setting the variable org-use-property-inheritance. It may be set to t to make all properties inherited from the parent, to a list of properties that should be inherited, or to a regular expression that matches inherited properties. If a property has the value nil, this is interpreted as an explicit undefine of the property, so that inheritance search will stop at this value and return nil.

Org mode has a few properties for which inheritance is hard-coded, at least for the special applications for which they are used:

The :COLUMNS: property defines the format of column view (see Column view). It is inherited in the sense that the level where a :COLUMNS: property is defined is used as the starting point for a column view table, independently of the location in the subtree from where columns view is turned on.
For agenda view, a category set through a :CATEGORY: property applies to the entire subtree.
For archiving, the :ARCHIVE: property may define the archive location for the entire subtree (see Moving subtrees).
The LOGGING property may define logging settings for an entry or a subtree (see Tracking TODO state changes).

Next: , Previous: Property inheritance, Up: Properties and Columns

7.5 Column view

A great way to view and edit properties in an outline tree is column view. In column view, each outline node is turned into a table row. Columns in this table provide access to properties of the entries. Org mode implements columns by overlaying a tabular structure over the headline of each item. While the headlines have been turned into a table row, you can still change the visibility of the outline tree. For example, you get a compact table by switching to CONTENTS view (S-<TAB> S-<TAB>, or simply c while column view is active), but you can still open, read, and edit the entry below each headline. Or, you can switch to column view after executing a sparse tree command and in this way get a table only for the selected items. Column view also works in agenda buffers (see Agenda Views) where queries have collected selected items, possibly from a number of files.

Next: , Previous: Column view, Up: Column view

7.5.1 Defining columns

Setting up a column view first requires defining the columns. This is done by defining a column format line.

Next: , Previous: Defining columns, Up: Defining columns Scope of column definitions

To define a column format for an entire file, use a line like


To specify a format that only applies to a specific tree, add a :COLUMNS: property to the top node of that tree, for example:

     ** Top node for columns view

If a :COLUMNS: property is present in an entry, it defines columns for the entry itself, and for the entire subtree below it. Since the column definition is part of the hierarchical structure of the document, you can define columns on level 1 that are general enough for all sublevels, and more specific columns further down, when you edit a deeper part of the tree.

Previous: Scope of column definitions, Up: Defining columns Column attributes

A column definition sets the attributes of a column. The general definition looks like this:


Except for the percent sign and the property name, all items are optional. The individual parts have the following meaning:

     width           An integer specifying the width of the column in characters.
                     If omitted, the width will be determined automatically.
     property        The property that should be edited in this column.
                     Special properties representing meta data are allowed here
                     as well (see Special properties)
     title           The header text for the column.  If omitted, the property
                     name is used.
     {summary-type}  The summary type.  If specified, the column values for
                     parent nodes are computed from the children.
                     Supported summary types are:
                     {+}       Sum numbers in this column.
                     {+;%.1f}  Like ‘+’, but format result with ‘%.1f’.
                     {$}       Currency, short for ‘+;%.2f’.
                     {:}       Sum times, HH:MM, plain numbers are hours.
                     {X}       Checkbox status, ‘[X]’ if all children are ‘[X]’.
                     {X/}      Checkbox status, ‘[n/m]’.
                     {X%}      Checkbox status, ‘[n%]’.
                     {min}     Smallest number in column.
                     {max}     Largest number.
                     {mean}    Arithmetic mean of numbers.
                     {:min}    Smallest time value in column.
                     {:max}    Largest time value.
                     {:mean}   Arithmetic mean of time values.
                     {@min}    Minimum age (in days/hours/mins/seconds).
                     {@max}    Maximum age (in days/hours/mins/seconds).
                     {@mean}   Arithmetic mean of ages (in days/hours/mins/seconds).
                     {est+}    Add low-high estimates.

Be aware that you can only have one summary type for any property you include. Subsequent columns referencing the same property will all display the same summary information.

The est+ summary type requires further explanation. It is used for combining estimates, expressed as low-high ranges. For example, instead of estimating a particular task will take 5 days, you might estimate it as 5–6 days if you're fairly confident you know how much work is required, or 1–10 days if you don't really know what needs to be done. Both ranges average at 5.5 days, but the first represents a more predictable delivery.

When combining a set of such estimates, simply adding the lows and highs produces an unrealistically wide result. Instead, est+ adds the statistical mean and variance of the sub-tasks, generating a final estimate from the sum. For example, suppose you had ten tasks, each of which was estimated at 0.5 to 2 days of work. Straight addition produces an estimate of 5 to 20 days, representing what to expect if everything goes either extremely well or extremely poorly. In contrast, est+ estimates the full job more realistically, at 10–15 days.

Numbers are right-aligned when a format specifier with an explicit width like %5d or %5.1f is used.

Here is an example for a complete columns definition, along with allowed values.

     :COLUMNS:  %25ITEM %9Approved(Approved?){X} %Owner %11Status \65
                        %10Time_Estimate{:} %CLOCKSUM %CLOCKSUM_T
     :Owner_ALL:    Tammy Mark Karl Lisa Don
     :Status_ALL:   "In progress" "Not started yet" "Finished" ""
     :Approved_ALL: "[ ]" "[X]"

The first column, ‘%25ITEM’, means the first 25 characters of the item itself, i.e., of the headline. You probably always should start the column definition with the ‘ITEM’ specifier. The other specifiers create columns ‘Owner’ with a list of names as allowed values, for ‘Status’ with four different possible values, and for a checkbox field ‘Approved’. When no width is given after the ‘%’ character, the column will be exactly as wide as it needs to be in order to fully display all values. The ‘Approved’ column does have a modified title (‘Approved?’, with a question mark). Summaries will be created for the ‘Time_Estimate’ column by adding time duration expressions like HH:MM, and for the ‘Approved’ column, by providing an ‘[X]’ status if all children have been checked. The ‘CLOCKSUM’ and ‘CLOCKSUM_T’ columns are special, they lists the sums of CLOCK intervals in the subtree, either for all clocks or just for today.

Next: , Previous: Defining columns, Up: Column view

7.5.2 Using column view

Turning column view on and off

C-c C-x C-c     (org-columns)
Turn on column view. If the cursor is before the first headline in the file, column view is turned on for the entire file, using the #+COLUMNS definition. If the cursor is somewhere inside the outline, this command searches the hierarchy, up from point, for a :COLUMNS: property that defines a format. When one is found, the column view table is established for the tree starting at the entry that contains the :COLUMNS: property. If no such property is found, the format is taken from the #+COLUMNS line or from the variable org-columns-default-format, and column view is established for the current entry and its subtree.
r     (org-columns-redo)
Recreate the column view, to include recent changes made in the buffer.
g     (org-columns-redo)
Same as r.
q     (org-columns-quit)
Exit column view.
Editing values
<left> <right> <up> <down>
Move through the column view from field to field.
Switch to the next/previous allowed value of the field. For this, you have to have specified allowed values for a property.
Directly select the Nth allowed value, 0 selects the 10th value.
n     (org-columns-next-allowed-value)
p     (org-columns-previous-allowed-value)
Same as S-<left>/<right>
e     (org-columns-edit-value)
Edit the property at point. For the special properties, this will invoke the same interface that you normally use to change that property. For example, when editing a TAGS property, the tag completion or fast selection interface will pop up.
C-c C-c     (org-columns-set-tags-or-toggle)
When there is a checkbox at point, toggle it.
v     (org-columns-show-value)
View the full value of this property. This is useful if the width of the column is smaller than that of the value.
a     (org-columns-edit-allowed)
Edit the list of allowed values for this property. If the list is found in the hierarchy, the modified values is stored there. If no list is found, the new value is stored in the first entry that is part of the current column view.
Modifying the table structure

<     (org-columns-narrow)
>     (org-columns-widen)
Make the column narrower/wider by one character.
S-M-<right>     (org-columns-new)
Insert a new column, to the left of the current column.
S-M-<left>     (org-columns-delete)
Delete the current column.

Previous: Using column view, Up: Column view

7.5.3 Capturing column view

Since column view is just an overlay over a buffer, it cannot be exported or printed directly. If you want to capture a column view, use a columnview dynamic block (see Dynamic blocks). The frame of this block looks like this:

     * The column view
     #+BEGIN: columnview :hlines 1 :id "label"

This dynamic block has the following parameters:

This is the most important parameter. Column view is a feature that is often localized to a certain (sub)tree, and the capture block might be at a different location in the file. To identify the tree whose view to capture, you can use 4 values:
          local     use the tree in which the capture block is located
          global    make a global view, including all headings in the file
                    run column view at the top of this file
          "ID"      call column view in the tree that has an :ID:
                    property with the value label.  You can use
                    M-x org-id-copy RET to create a globally unique ID for
                    the current entry and copy it to the kill-ring.

When t, insert an hline after every line. When a number N, insert an hline before each headline with level <= N.
When set to t, force column groups to get vertical lines.
When set to a number, don't capture entries below this level.
When set to t, skip rows where the only non-empty specifier of the column view is ITEM.

The following commands insert or update the dynamic block:

C-c C-x i     (org-insert-columns-dblock)
Insert a dynamic block capturing a column view. You will be prompted for the scope or ID of the view.
C-c C-c or C-c C-x C-u     (org-dblock-update)
Update dynamic block at point. The cursor needs to be in the #+BEGIN line of the dynamic block.
C-u C-c C-x C-u     (org-update-all-dblocks)
Update all dynamic blocks (see Dynamic blocks). This is useful if you have several clock table blocks, column-capturing blocks or other dynamic blocks in a buffer.

You can add formulas to the column view table and you may add plotting instructions in front of the table—these will survive an update of the block. If there is a #+TBLFM: after the table, the table will actually be recalculated automatically after an update.

An alternative way to capture and process property values into a table is provided by Eric Schulte's org-collector.el which is a contributed package66. It provides a general API to collect properties from entries in a certain scope, and arbitrary Lisp expressions to process these values before inserting them into a table or a dynamic block.

Previous: Column view, Up: Properties and Columns

7.6 The Property API

There is a full API for accessing and changing properties. This API can be used by Emacs Lisp programs to work with properties and to implement features based on them. For more information see Using the property API.

Next: , Previous: Properties and Columns, Up: Top

8 Dates and times

To assist project planning, TODO items can be labeled with a date and/or a time. The specially formatted string carrying the date and time information is called a timestamp in Org mode. This may be a little confusing because timestamp is often used as indicating when something was created or last changed. However, in Org mode this term is used in a much wider sense.

Next: , Previous: Dates and Times, Up: Dates and Times

8.1 Timestamps, deadlines, and scheduling

A timestamp is a specification of a date (possibly with a time or a range of times) in a special format, either ‘<2003-09-16 Tue>67 or ‘<2003-09-16 Tue 09:39>’ or ‘<2003-09-16 Tue 12:00-12:30>68. A timestamp can appear anywhere in the headline or body of an Org tree entry. Its presence causes entries to be shown on specific dates in the agenda (see Weekly/daily agenda). We distinguish:

Plain timestamp; Event; Appointment
A simple timestamp just assigns a date/time to an item. This is just like writing down an appointment or event in a paper agenda. In the timeline and agenda displays, the headline of an entry associated with a plain timestamp will be shown exactly on that date.
          * Meet Peter at the movies
            <2006-11-01 Wed 19:15>
          * Discussion on climate change
            <2006-11-02 Thu 20:00-22:00>

Timestamp with repeater interval
A timestamp may contain a repeater interval, indicating that it applies not only on the given date, but again and again after a certain interval of N days (d), weeks (w), months (m), or years (y). The following will show up in the agenda every Wednesday:
          * Pick up Sam at school
            <2007-05-16 Wed 12:30 +1w>

Diary-style sexp entries
For more complex date specifications, Org mode supports using the special sexp diary entries implemented in the Emacs calendar/diary package69. For example with optional time
          * 22:00-23:00 The nerd meeting on every 2nd Thursday of the month
            <%%(diary-float t 4 2)>

Time/Date range
Two timestamps connected by ‘--’ denote a range. The headline will be shown on the first and last day of the range, and on any dates that are displayed and fall in the range. Here is an example:
          ** Meeting in Amsterdam
             <2004-08-23 Mon>--<2004-08-26 Thu>

Inactive timestamp
Just like a plain timestamp, but with square brackets instead of angular ones. These timestamps are inactive in the sense that they do not trigger an entry to show up in the agenda.
          * Gillian comes late for the fifth time
            [2006-11-01 Wed]

Next: , Previous: Timestamps, Up: Dates and Times

8.2 Creating timestamps

For Org mode to recognize timestamps, they need to be in the specific format. All commands listed below produce timestamps in the correct format.

C-c .     (org-time-stamp)
Prompt for a date and insert a corresponding timestamp. When the cursor is at an existing timestamp in the buffer, the command is used to modify this timestamp instead of inserting a new one. When this command is used twice in succession, a time range is inserted.
C-c !     (org-time-stamp-inactive)
Like C-c ., but insert an inactive timestamp that will not cause an agenda entry.
C-u C-c .
C-u C-c !
Like C-c . and C-c !, but use the alternative format which contains date and time. The default time can be rounded to multiples of 5 minutes, see the option org-time-stamp-rounding-minutes.
C-c C-c
Normalize timestamp, insert/fix day name if missing or wrong.
C-c <     (org-date-from-calendar)
Insert a timestamp corresponding to the cursor date in the Calendar.
C-c >     (org-goto-calendar)
Access the Emacs calendar for the current date. If there is a timestamp in the current line, go to the corresponding date instead.
C-c C-o     (org-open-at-point)
Access the agenda for the date given by the timestamp or -range at point (see Weekly/daily agenda).
S-<left>     (org-timestamp-down-day)
S-<right>     (org-timestamp-up-day)
Change date at cursor by one day. These key bindings conflict with shift-selection and related modes (see Conflicts).
S-<up>     (org-timestamp-up)
S-<down>     (org-timestamp-down-down)
Change the item under the cursor in a timestamp. The cursor can be on a year, month, day, hour or minute. When the timestamp contains a time range like ‘15:30-16:30’, modifying the first time will also shift the second, shifting the time block with constant length. To change the length, modify the second time. Note that if the cursor is in a headline and not at a timestamp, these same keys modify the priority of an item. (see Priorities). The key bindings also conflict with shift-selection and related modes (see Conflicts).
C-c C-y     (org-evaluate-time-range)
Evaluate a time range by computing the difference between start and end. With a prefix argument, insert result after the time range (in a table: into the following column).

Next: , Previous: Creating timestamps, Up: Creating timestamps

8.2.1 The date/time prompt

When Org mode prompts for a date/time, the default is shown in default date/time format, and the prompt therefore seems to ask for a specific format. But it will in fact accept date/time information in a variety of formats. Generally, the information should start at the beginning of the string. Org mode will find whatever information is in there and derive anything you have not specified from the default date and time. The default is usually the current date and time, but when modifying an existing timestamp, or when entering the second stamp of a range, it is taken from the stamp in the buffer. When filling in information, Org mode assumes that most of the time you will want to enter a date in the future: if you omit the month/year and the given day/month is before today, it will assume that you mean a future date70. If the date has been automatically shifted into the future, the time prompt will show this with ‘(=>F).

For example, let's assume that today is June 13, 2006. Here is how various inputs will be interpreted, the items filled in by Org mode are in bold.

     3-2-5         ⇒ 2003-02-05
     2/5/3         ⇒ 2003-02-05
     14            ⇒ 2006-06-14
     12            ⇒ 2006-07-12
     2/5           ⇒ 2007-02-05
     Fri           ⇒ nearest Friday after the default date
     sep 15        ⇒ 2006-09-15
     feb 15        ⇒ 2007-02-15
     sep 12 9      ⇒ 2009-09-12
     12:45         ⇒ 2006-06-13 12:45
     22 sept 0:34  ⇒ 2006-09-22 00:34
     w4            ⇒ ISO week for of the current year 2006
     2012 w4 fri   ⇒ Friday of ISO week 4 in 2012
     2012-w04-5    ⇒ Same as above

Furthermore you can specify a relative date by giving, as the first thing in the input: a plus/minus sign, a number and a letter ([hdwmy]) to indicate change in hours, days, weeks, months, or years. With a single plus or minus, the date is always relative to today. With a double plus or minus, it is relative to the default date. If instead of a single letter, you use the abbreviation of day name, the date will be the Nth such day, e.g.:

     +0            ⇒ today
     .             ⇒ today
     +4d           ⇒ four days from today
     +4            ⇒ same as above
     +2w           ⇒ two weeks from today
     ++5           ⇒ five days from default date
     +2tue         ⇒ second Tuesday from now
     -wed          ⇒ last Wednesday

The function understands English month and weekday abbreviations. If you want to use unabbreviated names and/or other languages, configure the variables parse-time-months and parse-time-weekdays.

Not all dates can be represented in a given Emacs implementation. By default Org mode forces dates into the compatibility range 1970–2037 which works on all Emacs implementations. If you want to use dates outside of this range, read the docstring of the variable org-read-date-force-compatible-dates.

You can specify a time range by giving start and end times or by giving a start time and a duration (in HH:MM format). Use one or two dash(es) as the separator in the former case and use '+' as the separator in the latter case, e.g.:

     11am-1:15pm    ⇒ 11:00-13:15
     11am--1:15pm   ⇒ same as above
     11am+2:15      ⇒ same as above

Parallel to the minibuffer prompt, a calendar is popped up71. When you exit the date prompt, either by clicking on a date in the calendar, or by pressing <RET>, the date selected in the calendar will be combined with the information entered at the prompt. You can control the calendar fully from the minibuffer:

     <RET>           Choose date at cursor in calendar.
     mouse-1        Select date by clicking on it.
     S-<right>/<left>     One day forward/backward.
     S-<down>/<up>     One week forward/backward.
     M-S-<right>/<left>   One month forward/backward.
     > / <          Scroll calendar forward/backward by one month.
     M-v / C-v      Scroll calendar forward/backward by 3 months.

The actions of the date/time prompt may seem complex, but I assure you they will grow on you, and you will start getting annoyed by pretty much any other way of entering a date/time out there. To help you understand what is going on, the current interpretation of your input will be displayed live in the minibuffer72.

Previous: The date/time prompt, Up: Creating timestamps

8.2.2 Custom time format

Org mode uses the standard ISO notation for dates and times as it is defined in ISO 8601. If you cannot get used to this and require another representation of date and time to keep you happy, you can get it by customizing the options org-display-custom-times and org-time-stamp-custom-formats.

C-c C-x C-t     (org-toggle-time-stamp-overlays)
Toggle the display of custom formats for dates and times.

Org mode needs the default format for scanning, so the custom date/time format does not replace the default format—instead it is put over the default format using text properties. This has the following consequences:

Next: , Previous: Creating timestamps, Up: Dates and Times

8.3 Deadlines and scheduling

A timestamp may be preceded by special keywords to facilitate planning:

Meaning: the task (most likely a TODO item, though not necessarily) is supposed to be finished on that date.

On the deadline date, the task will be listed in the agenda. In addition, the agenda for today will carry a warning about the approaching or missed deadline, starting org-deadline-warning-days before the due date, and continuing until the entry is marked DONE. An example:

          *** TODO write article about the Earth for the Guide
              DEADLINE: <2004-02-29 Sun>
              The editor in charge is [[bbdb:Ford Prefect]]

You can specify a different lead time for warnings for a specific deadlines using the following syntax. Here is an example with a warning period of 5 days DEADLINE: <2004-02-29 Sun -5d>. This warning is deactivated if the task get scheduled and you set org-agenda-skip-deadline-prewarning-if-scheduled to t.

Meaning: you are planning to start working on that task on the given date.

The headline will be listed under the given date73. In addition, a reminder that the scheduled date has passed will be present in the compilation for today, until the entry is marked DONE, i.e., the task will automatically be forwarded until completed.

          *** TODO Call Trillian for a date on New Years Eve.
              SCHEDULED: <2004-12-25 Sat>

If you want to delay the display of this task in the agenda, use SCHEDULED: <2004-12-25 Sat -2d>: the task is still scheduled on the 25th but will appear two days later. In case the task contains a repeater, the delay is considered to affect all occurrences; if you want the delay to only affect the first scheduled occurrence of the task, use --2d instead. See org-scheduled-delay-days and org-agenda-skip-scheduled-delay-if-deadline for details on how to control this globally or per agenda.

Important: Scheduling an item in Org mode should not be understood in the same way that we understand scheduling a meeting. Setting a date for a meeting is just a simple appointment, you should mark this entry with a simple plain timestamp, to get this item shown on the date where it applies. This is a frequent misunderstanding by Org users. In Org mode, scheduling means setting a date when you want to start working on an action item.

You may use timestamps with repeaters in scheduling and deadline entries. Org mode will issue early and late warnings based on the assumption that the timestamp represents the nearest instance of the repeater. However, the use of diary sexp entries like <%%(diary-float t 42)> in scheduling and deadline timestamps is limited. Org mode does not know enough about the internals of each sexp function to issue early and late warnings. However, it will show the item on each day where the sexp entry matches.

Next: , Previous: Deadlines and scheduling, Up: Deadlines and scheduling

8.3.1 Inserting deadlines or schedules

The following commands allow you to quickly insert74 a deadline or to schedule an item:

C-c C-d     (org-deadline)
Insert ‘DEADLINE’ keyword along with a stamp. The insertion will happen in the line directly following the headline. Any CLOSED timestamp will be removed. When called with a prefix arg, an existing deadline will be removed from the entry. Depending on the variable org-log-redeadline75, a note will be taken when changing an existing deadline.

C-c C-s     (org-schedule)
Insert ‘SCHEDULED’ keyword along with a stamp. The insertion will happen in the line directly following the headline. Any CLOSED timestamp will be removed. When called with a prefix argument, remove the scheduling date from the entry. Depending on the variable org-log-reschedule76, a note will be taken when changing an existing scheduling time.
C-c C-x C-k     (org-mark-entry-for-agenda-action)
Mark the current entry for agenda action. After you have marked the entry like this, you can open the agenda or the calendar to find an appropriate date. With the cursor on the selected date, press k s or k d to schedule the marked item.
C-c / d     (org-check-deadlines)
Create a sparse tree with all deadlines that are either past-due, or which will become due within org-deadline-warning-days. With C-u prefix, show all deadlines in the file. With a numeric prefix, check that many days. For example, C-1 C-c / d shows all deadlines due tomorrow.
C-c / b     (org-check-before-date)
Sparse tree for deadlines and scheduled items before a given date.
C-c / a     (org-check-after-date)
Sparse tree for deadlines and scheduled items after a given date.

Note that org-schedule and org-deadline supports setting the date by indicating a relative time: e.g., +1d will set the date to the next day after today, and –1w will set the date to the previous week before any current timestamp.

Previous: Inserting deadline/schedule, Up: Deadlines and scheduling

8.3.2 Repeated tasks

Some tasks need to be repeated again and again. Org mode helps to organize such tasks using a so-called repeater in a DEADLINE, SCHEDULED, or plain timestamp. In the following example

     ** TODO Pay the rent
        DEADLINE: <2005-10-01 Sat +1m>

the +1m is a repeater; the intended interpretation is that the task has a deadline on <2005-10-01> and repeats itself every (one) month starting from that time. You can use yearly, monthly, weekly, daily and hourly repeat cookies by using the y/w/m/d/h letters. If you need both a repeater and a special warning period in a deadline entry, the repeater should come first and the warning period last: DEADLINE: <2005-10-01 Sat +1m -3d>.

Deadlines and scheduled items produce entries in the agenda when they are over-due, so it is important to be able to mark such an entry as completed once you have done so. When you mark a DEADLINE or a SCHEDULE with the TODO keyword DONE, it will no longer produce entries in the agenda. The problem with this is, however, that then also the next instance of the repeated entry will not be active. Org mode deals with this in the following way: When you try to mark such an entry DONE (using C-c C-t), it will shift the base date of the repeating timestamp by the repeater interval, and immediately set the entry state back to TODO77. In the example above, setting the state to DONE would actually switch the date like this:

     ** TODO Pay the rent
        DEADLINE: <2005-11-01 Tue +1m>

A timestamp78 will be added under the deadline, to keep a record that you actually acted on the previous instance of this deadline.

As a consequence of shifting the base date, this entry will no longer be visible in the agenda when checking past dates, but all future instances will be visible.

With the ‘+1m’ cookie, the date shift will always be exactly one month. So if you have not paid the rent for three months, marking this entry DONE will still keep it as an overdue deadline. Depending on the task, this may not be the best way to handle it. For example, if you forgot to call your father for 3 weeks, it does not make sense to call him 3 times in a single day to make up for it. Finally, there are tasks like changing batteries which should always repeat a certain time after the last time you did it. For these tasks, Org mode has special repeaters ‘++’ and ‘.+’. For example:

     ** TODO Call Father
        DEADLINE: <2008-02-10 Sun ++1w>
        Marking this DONE will shift the date by at least one week,
        but also by as many weeks as it takes to get this date into
        the future.  However, it stays on a Sunday, even if you called
        and marked it done on Saturday.
     ** TODO Check the batteries in the smoke detectors
        DEADLINE: <2005-11-01 Tue .+1m>
        Marking this DONE will shift the date to one month after

You may have both scheduling and deadline information for a specific task. If the repeater is set for the scheduling information only, you probably want the repeater to be ignored after the deadline. If so, set the variable org-agenda-skip-scheduled-if-deadline-is-shown to repeated-after-deadline. If you want both scheduling and deadline information to repeat after the same interval, set the same repeater for both timestamps.

An alternative to using a repeater is to create a number of copies of a task subtree, with dates shifted in each copy. The command C-c C-x c was created for this purpose, it is described in Structure editing.

Next: , Previous: Deadlines and scheduling, Up: Dates and Times

8.4 Clocking work time

Org mode allows you to clock the time you spend on specific tasks in a project. When you start working on an item, you can start the clock. When you stop working on that task, or when you mark the task done, the clock is stopped and the corresponding time interval is recorded. It also computes the total time spent on each subtree79 of a project. And it remembers a history or tasks recently clocked, to that you can jump quickly between a number of tasks absorbing your time.

To save the clock history across Emacs sessions, use

     (setq org-clock-persist 'history)

When you clock into a new task after resuming Emacs, the incomplete clock80 will be found (see Resolving idle time) and you will be prompted about what to do with it.

Next: , Previous: Clocking work time, Up: Clocking work time

8.4.1 Clocking commands

C-c C-x C-i     (org-clock-in)
Start the clock on the current item (clock-in). This inserts the CLOCK keyword together with a timestamp. If this is not the first clocking of this item, the multiple CLOCK lines will be wrapped into a :LOGBOOK: drawer (see also the variable org-clock-into-drawer). You can also overrule the setting of this variable for a subtree by setting a CLOCK_INTO_DRAWER or LOG_INTO_DRAWER property. When called with a C-u prefix argument, select the task from a list of recently clocked tasks. With two C-u C-u prefixes, clock into the task at point and mark it as the default task; the default task will then always be available with letter d when selecting a clocking task. With three C-u C-u C-u prefixes, force continuous clocking by starting the clock when the last clock stopped.
While the clock is running, the current clocking time is shown in the mode line, along with the title of the task. The clock time shown will be all time ever clocked for this task and its children. If the task has an effort estimate (see Effort estimates), the mode line displays the current clocking time against it81 If the task is a repeating one (see Repeated tasks), only the time since the last reset of the task 82 will be shown. More control over what time is shown can be exercised with the CLOCK_MODELINE_TOTAL property. It may have the values current to show only the current clocking instance, today to show all time clocked on this tasks today (see also the variable org-extend-today-until), all to include all time, or auto which is the default83.
Clicking with mouse-1 onto the mode line entry will pop up a menu with clocking options.
C-c C-x C-o     (org-clock-out)
Stop the clock (clock-out). This inserts another timestamp at the same location where the clock was last started. It also directly computes the resulting time in inserts it after the time range as ‘=> HH:MM’. See the variable org-log-note-clock-out for the possibility to record an additional note together with the clock-out timestamp84.
C-c C-x C-x     (org-clock-in-last)
Reclock the last clocked task. With one C-u prefix argument, select the task from the clock history. With two C-u prefixes, force continuous clocking by starting the clock when the last clock stopped.
C-c C-x C-e     (org-clock-modify-effort-estimate)
Update the effort estimate for the current clock task.
C-c C-c or C-c C-y     (org-evaluate-time-range)
Recompute the time interval after changing one of the timestamps. This is only necessary if you edit the timestamps directly. If you change them with S-<cursor> keys, the update is automatic.
C-S-<up/down>     (org-clock-timestamps-up/down)
On CLOCK log lines, increase/decrease both timestamps so that the clock duration keeps the same.
S-M-<up/down>     (org-timestamp-up/down)
On CLOCK log lines, increase/decrease the timestamp at point and the one of the previous (or the next clock) timestamp by the same duration. For example, if you hit S-M-<up> to increase a clocked-out timestamp by five minutes, then the clocked-in timestamp of the next clock will be increased by five minutes.
C-c C-t     (org-todo)
Changing the TODO state of an item to DONE automatically stops the clock if it is running in this same item.
C-c C-x C-q     (org-clock-cancel)
Cancel the current clock. This is useful if a clock was started by mistake, or if you ended up working on something else.
C-c C-x C-j     (org-clock-goto)
Jump to the headline of the currently clocked in task. With a C-u prefix arg, select the target task from a list of recently clocked tasks.
C-c C-x C-d     (org-clock-display)
Display time summaries for each subtree in the current buffer. This puts overlays at the end of each headline, showing the total time recorded under that heading, including the time of any subheadings. You can use visibility cycling to study the tree, but the overlays disappear when you change the buffer (see variable org-remove-highlights-with-change) or press C-c C-c.

The l key may be used in the timeline (see Timeline) and in the agenda (see Weekly/daily agenda) to show which tasks have been worked on or closed during a day.

Important: note that both org-clock-out and org-clock-in-last can have a global keybinding and will not modify the window disposition.

Next: , Previous: Clocking commands, Up: Clocking work time

8.4.2 The clock table

Org mode can produce quite complex reports based on the time clocking information. Such a report is called a clock table, because it is formatted as one or several Org tables.

C-c C-x C-r     (org-clock-report)
Insert a dynamic block (see Dynamic blocks) containing a clock report as an Org mode table into the current file. When the cursor is at an existing clock table, just update it. When called with a prefix argument, jump to the first clock report in the current document and update it. The clock table always includes also trees with :ARCHIVE: tag.
C-c C-c or C-c C-x C-u     (org-dblock-update)
Update dynamic block at point. The cursor needs to be in the #+BEGIN line of the dynamic block.
C-u C-c C-x C-u
Update all dynamic blocks (see Dynamic blocks). This is useful if you have several clock table blocks in a buffer.
S-<right>     (org-clocktable-try-shift)
Shift the current :block interval and update the table. The cursor needs to be in the #+BEGIN: clocktable line for this command. If :block is today, it will be shifted to today-1 etc.

Here is an example of the frame for a clock table as it is inserted into the buffer with the C-c C-x C-r command:

     #+BEGIN: clocktable :maxlevel 2 :emphasize nil :scope file
     #+END: clocktable

The ‘BEGIN’ line and specify a number of options to define the scope, structure, and formatting of the report. Defaults for all these options can be configured in the variable org-clocktable-defaults.

First there are options that determine which clock entries are to be selected:

     :maxlevel    Maximum level depth to which times are listed in the table.
                  Clocks at deeper levels will be summed into the upper level.
     :scope       The scope to consider.  This can be any of the following:
                  nil        the current buffer or narrowed region
                  file       the full current buffer
                  subtree    the subtree where the clocktable is located
                  treeN      the surrounding level N tree, for example tree3
                  tree       the surrounding level 1 tree
                  agenda     all agenda files
                  ("file"..) scan these files
                  file-with-archives    current file and its archives
                  agenda-with-archives  all agenda files, including archives
     :block       The time block to consider.  This block is specified either
                  absolute, or relative to the current time and may be any of
                  these formats:
                  2007-12-31    New year eve 2007
                  2007-12       December 2007
                  2007-W50      ISO-week 50 in 2007
                  2007-Q2       2nd quarter in 2007
                  2007          the year 2007
                  today, yesterday, today-N          a relative day
                  thisweek, lastweek, thisweek-N     a relative week
                  thismonth, lastmonth, thismonth-N  a relative month
                  thisyear, lastyear, thisyear-N     a relative year
                  Use S-<left>/<right> keys to shift the time interval.
     :tstart      A time string specifying when to start considering times.
                  Relative times like "<-2w>" can also be used.  See
                  Matching tags and properties for relative time syntax.
     :tend        A time string specifying when to stop considering times.
                  Relative times like "<now>" can also be used.  See
                  Matching tags and properties for relative time syntax.
     :wstart      The starting day of the week.  The default is 1 for monday.
     :mstart      The starting day of the month.  The default 1 is for the first
                  day of the month.
     :step        week or day, to split the table into chunks.
                  To use this, :block or :tstart, :tend are needed.
     :stepskip0   Do not show steps that have zero time.
     :fileskip0   Do not show table sections from files which did not contribute.
     :tags        A tags match to select entries that should contribute.  See
                  Matching tags and properties for the match syntax.

Then there are options which determine the formatting of the table. There options are interpreted by the function org-clocktable-write-default, but you can specify your own function using the :formatter parameter.

     :emphasize   When t, emphasize level one and level two items.
     :lang        Language85 to use for descriptive cells like "Task".
     :link        Link the item headlines in the table to their origins.
     :narrow      An integer to limit the width of the headline column in
                  the org table.  If you write it like ‘50!’, then the
                  headline will also be shortened in export.
     :indent      Indent each headline field according to its level.
     :tcolumns    Number of columns to be used for times.  If this is smaller
                  than :maxlevel, lower levels will be lumped into one column.
     :level       Should a level number column be included?
     :compact     Abbreviation for :level nil :indent t :narrow 40! :tcolumns 1
                  All are overwritten except if there is an explicit :narrow
     :timestamp   A timestamp for the entry, when available.  Look for SCHEDULED,
                  DEADLINE, TIMESTAMP and TIMESTAMP_IA, in this order.
     :properties  List of properties that should be shown in the table.  Each
                  property will get its own column.
     :inherit-props When this flag is t, the values for :properties will be inherited.
     :formula     Content of a #+TBLFM line to be added and evaluated.
                  As a special case, ‘:formula %’ adds a column with % time.
                  If you do not specify a formula here, any existing formula
                  below the clock table will survive updates and be evaluated.
     :formatter   A function to format clock data and insert it into the buffer.

To get a clock summary of the current level 1 tree, for the current day, you could write

     #+BEGIN: clocktable :maxlevel 2 :block today :scope tree1 :link t
     #+END: clocktable

and to use a specific time range you could write86

     #+BEGIN: clocktable :tstart "<2006-08-10 Thu 10:00>"
                         :tend "<2006-08-10 Thu 12:00>"
     #+END: clocktable

A range starting a week ago and ending right now could be written as

     #+BEGIN: clocktable :tstart "<-1w>" :tend "<now>"
     #+END: clocktable

A summary of the current subtree with % times would be

     #+BEGIN: clocktable :scope subtree :link t :formula %
     #+END: clocktable

A horizontally compact representation of everything clocked during last week would be

     #+BEGIN: clocktable :scope agenda :block lastweek :compact t
     #+END: clocktable

Previous: The clock table, Up: Clocking work time

8.4.3 Resolving idle time and continuous clocking

Resolving idle time

If you clock in on a work item, and then walk away from your computer—perhaps to take a phone call—you often need to “resolve” the time you were away by either subtracting it from the current clock, or applying it to another one.

By customizing the variable org-clock-idle-time to some integer, such as 10 or 15, Emacs can alert you when you get back to your computer after being idle for that many minutes87, and ask what you want to do with the idle time. There will be a question waiting for you when you get back, indicating how much idle time has passed (constantly updated with the current amount), as well as a set of choices to correct the discrepancy:

To keep some or all of the minutes and stay clocked in, press k. Org will ask how many of the minutes to keep. Press <RET> to keep them all, effectively changing nothing, or enter a number to keep that many minutes.
If you use the shift key and press K, it will keep however many minutes you request and then immediately clock out of that task. If you keep all of the minutes, this is the same as just clocking out of the current task.
To keep none of the minutes, use s to subtract all the away time from the clock, and then check back in from the moment you returned.
To keep none of the minutes and just clock out at the start of the away time, use the shift key and press S. Remember that using shift will always leave you clocked out, no matter which option you choose.
To cancel the clock altogether, use C. Note that if instead of canceling you subtract the away time, and the resulting clock amount is less than a minute, the clock will still be canceled rather than clutter up the log with an empty entry.

What if you subtracted those away minutes from the current clock, and now want to apply them to a new clock? Simply clock in to any task immediately after the subtraction. Org will notice that you have subtracted time “on the books”, so to speak, and will ask if you want to apply those minutes to the next task you clock in on.

There is one other instance when this clock resolution magic occurs. Say you were clocked in and hacking away, and suddenly your cat chased a mouse who scared a hamster that crashed into your UPS's power button! You suddenly lose all your buffers, but thanks to auto-save you still have your recent Org mode changes, including your last clock in.

If you restart Emacs and clock into any task, Org will notice that you have a dangling clock which was never clocked out from your last session. Using that clock's starting time as the beginning of the unaccounted-for period, Org will ask how you want to resolve that time. The logic and behavior is identical to dealing with away time due to idleness; it is just happening due to a recovery event rather than a set amount of idle time.

You can also check all the files visited by your Org agenda for dangling clocks at any time using M-x org-resolve-clocks RET (or C-c C-x C-z).

Continuous clocking

You may want to start clocking from the time when you clocked out the previous task. To enable this systematically, set org-clock-continuously to t. Each time you clock in, Org retrieves the clock-out time of the last clocked entry for this session, and start the new clock from there.

If you only want this from time to time, use three universal prefix arguments with org-clock-in and two C-u C-u with org-clock-in-last.

Next: , Previous: Clocking work time, Up: Dates and Times

8.5 Effort estimates

If you want to plan your work in a very detailed way, or if you need to produce offers with quotations of the estimated work effort, you may want to assign effort estimates to entries. If you are also clocking your work, you may later want to compare the planned effort with the actual working time, a great way to improve planning estimates. Effort estimates are stored in a special property ‘Effort88. You can set the effort for an entry with the following commands:

C-c C-x e     (org-set-effort)
Set the effort estimate for the current entry. With a numeric prefix argument, set it to the Nth allowed value (see below). This command is also accessible from the agenda with the e key.
C-c C-x C-e     (org-clock-modify-effort-estimate)
Modify the effort estimate of the item currently being clocked.

Clearly the best way to work with effort estimates is through column view (see Column view). You should start by setting up discrete values for effort estimates, and a COLUMNS format that displays these values together with clock sums (if you want to clock your time). For a specific buffer you can use

     #+PROPERTY: Effort_ALL 0 0:10 0:30 1:00 2:00 3:00 4:00 5:00 6:00 7:00
     #+COLUMNS: %40ITEM(Task) %17Effort(Estimated Effort){:} %CLOCKSUM

or, even better, you can set up these values globally by customizing the variables org-global-properties and org-columns-default-format. In particular if you want to use this setup also in the agenda, a global setup may be advised.

The way to assign estimates to individual items is then to switch to column mode, and to use S-<right> and S-<left> to change the value. The values you enter will immediately be summed up in the hierarchy. In the column next to it, any clocked time will be displayed.

If you switch to column view in the daily/weekly agenda, the effort column will summarize the estimated work effort for each day89, and you can use this to find space in your schedule. To get an overview of the entire part of the day that is committed, you can set the option org-agenda-columns-add-appointments-to-effort-sum. The appointments on a day that take place over a specified time interval will then also be added to the load estimate of the day.

Effort estimates can be used in secondary agenda filtering that is triggered with the / key in the agenda (see Agenda commands). If you have these estimates defined consistently, two or three key presses will narrow down the list to stuff that fits into an available time slot.

Next: , Previous: Effort estimates, Up: Dates and Times

8.6 Taking notes with a relative timer

When taking notes during, for example, a meeting or a video viewing, it can be useful to have access to times relative to a starting time. Org provides such a relative timer and make it easy to create timed notes.

C-c C-x .     (org-timer)
Insert a relative time into the buffer. The first time you use this, the timer will be started. When called with a prefix argument, the timer is restarted.
C-c C-x -     (org-timer-item)
Insert a description list item with the current relative time. With a prefix argument, first reset the timer to 0.
M-<RET>     (org-insert-heading)
Once the timer list is started, you can also use M-<RET> to insert new timer items.
C-c C-x ,
Pause the timer, or continue it if it is already paused (org-timer-pause-or-continue).
C-u C-c C-x ,
Stop the timer. After this, you can only start a new timer, not continue the old one. This command also removes the timer from the mode line.
C-c C-x 0     (org-timer-start)
Reset the timer without inserting anything into the buffer. By default, the timer is reset to 0. When called with a C-u prefix, reset the timer to specific starting offset. The user is prompted for the offset, with a default taken from a timer string at point, if any, So this can be used to restart taking notes after a break in the process. When called with a double prefix argument C-u C-u, change all timer strings in the active region by a certain amount. This can be used to fix timer strings if the timer was not started at exactly the right moment.

Previous: Relative timer, Up: Dates and Times

8.7 Countdown timer

Calling org-timer-set-timer from an Org mode buffer runs a countdown timer. Use ; from agenda buffers, <C-c C-x ;> everywhere else.

org-timer-set-timer prompts the user for a duration and displays a countdown timer in the modeline. org-timer-default-timer sets the default countdown value. Giving a prefix numeric argument overrides this default value.

Next: , Previous: Dates and Times, Up: Top

9 Capture - Refile - Archive

An important part of any organization system is the ability to quickly capture new ideas and tasks, and to associate reference material with them. Org does this using a process called capture. It also can store files related to a task (attachments) in a special directory. Once in the system, tasks and projects need to be moved around. Moving completed project trees to an archive file keeps the system compact and fast.

Next: , Previous: Capture - Refile - Archive, Up: Capture - Refile - Archive

9.1 Capture

Capture lets you quickly store notes with little interruption of your work flow. Org's method for capturing new items is heavily inspired by John Wiegley excellent remember.el package. Up to version 6.36, Org used a special setup for remember.el, then replaced it with org-remember.el. As of version 8.0, org-remember.el has been completely replaced by org-capture.el.

If your configuration depends on org-remember.el, you need to update it and use the setup described below. To convert your org-remember-templates, run the command

     M-x org-capture-import-remember-templates RET

and then customize the new variable with M-x customize-variable org-capture-templates, check the result, and save the customization.

Next: , Previous: Capture, Up: Capture

9.1.1 Setting up capture

The following customization sets a default target file for notes, and defines a global key90 for capturing new material.

     (setq org-default-notes-file (concat org-directory "/"))
     (define-key global-map "\C-cc" 'org-capture)

Next: , Previous: Setting up capture, Up: Capture

9.1.2 Using capture

C-c c     (org-capture)
Call the command org-capture. Note that this keybinding is global and not active by default: you need to install it. If you have templates defined see Capture templates, it will offer these templates for selection or use a new Org outline node as the default template. It will insert the template into the target file and switch to an indirect buffer narrowed to this new node. You may then insert the information you want.

C-c C-c     (org-capture-finalize)
Once you have finished entering information into the capture buffer, C-c C-c will return you to the window configuration before the capture process, so that you can resume your work without further distraction. When called with a prefix arg, finalize and then jump to the captured item.

C-c C-w     (org-capture-refile)
Finalize the capture process by refiling (see Refile and copy) the note to a different place. Please realize that this is a normal refiling command that will be executed—so the cursor position at the moment you run this command is important. If you have inserted a tree with a parent and children, first move the cursor back to the parent. Any prefix argument given to this command will be passed on to the org-refile command.

C-c C-k     (org-capture-kill)
Abort the capture process and return to the previous state.

You can also call org-capture in a special way from the agenda, using the k c key combination. With this access, any timestamps inserted by the selected capture template will default to the cursor date in the agenda, rather than to the current date.

To find the locations of the last stored capture, use org-capture with prefix commands:

C-u C-c c
Visit the target location of a capture template. You get to select the template in the usual way.
C-u C-u C-c c
Visit the last stored capture item in its buffer.

You can also jump to the bookmark org-capture-last-stored, which will automatically be created unless you set org-capture-bookmark to nil.

To insert the capture at point in an Org buffer, call org-capture with a C-0 prefix argument.

Previous: Using capture, Up: Capture

9.1.3 Capture templates

You can use templates for different types of capture items, and for different target locations. The easiest way to create such templates is through the customize interface.

C-c c C
Customize the variable org-capture-templates.

Before we give the formal description of template definitions, let's look at an example. Say you would like to use one template to create general TODO entries, and you want to put these entries under the heading ‘Tasks’ in your file ~/org/ Also, a date tree in the file should capture journal entries. A possible configuration would look like:

     (setq org-capture-templates
      '(("t" "Todo" entry (file+headline "~/org/" "Tasks")
             "* TODO %?\n  %i\n  %a")
        ("j" "Journal" entry (file+datetree "~/org/")
             "* %?\nEntered on %U\n  %i\n  %a")))

If you then press C-c c t, Org will prepare the template for you like this:

     * TODO
       [[file:link to where you initiated capture]]

During expansion of the template, %a has been replaced by a link to the location from where you called the capture command. This can be extremely useful for deriving tasks from emails, for example. You fill in the task definition, press C-c C-c and Org returns you to the same place where you started the capture process.

To define special keys to capture to a particular template without going through the interactive template selection, you can create your key binding like this:

     (define-key global-map "\C-cx"
        (lambda () (interactive) (org-capture nil "x")))

Next: , Previous: Capture templates, Up: Capture templates Template elements

Now lets look at the elements of a template definition. Each entry in org-capture-templates is a list with the following items:

The keys that will select the template, as a string, characters only, for example "a" for a template to be selected with a single key, or "bt" for selection with two keys. When using several keys, keys using the same prefix key must be sequential in the list and preceded by a 2-element entry explaining the prefix key, for example
                   ("b" "Templates for marking stuff to buy")

If you do not define a template for the C key, this key will be used to open the customize buffer for this complex variable.

A short string describing the template, which will be shown during selection.
The type of entry, a symbol. Valid values are:
An Org mode node, with a headline. Will be filed as the child of the target entry or as a top-level entry. The target file should be an Org mode file.
A plain list item, placed in the first plain list at the target location. Again the target file should be an Org file.
A checkbox item. This only differs from the plain list item by the default template.
a new line in the first table at the target location. Where exactly the line will be inserted depends on the properties :prepend and :table-line-pos (see below).
Text to be inserted as it is.

Specification of where the captured item should be placed. In Org mode files, targets usually define a node. Entries will become children of this node. Other types will be added to the table or list in the body of this node. Most target specifications contain a file name. If that file name is the empty string, it defaults to org-default-notes-file. A file can also be given as a variable, function, or Emacs Lisp form.

Valid values are:

(file "path/to/file")
Text will be placed at the beginning or end of that file.
(id "id of existing org entry")
Filing as child of this entry, or in the body of the entry.
(file+headline "path/to/file" "node headline")
Fast configuration if the target heading is unique in the file.
(file+olp "path/to/file" "Level 1 heading" "Level 2" ...)
For non-unique headings, the full path is safer.
(file+regexp "path/to/file" "regexp to find location")
Use a regular expression to position the cursor.
(file+datetree "path/to/file")
Will create a heading in a date tree for today's date91.
(file+datetree+prompt "path/to/file")
Will create a heading in a date tree, but will prompt for the date.
(file+function "path/to/file" function-finding-location)
A function to find the right location in the file.
File to the entry that is currently being clocked.
(function function-finding-location)
Most general way, write your own function to find both file and location.

The template for creating the capture item. If you leave this empty, an appropriate default template will be used. Otherwise this is a string with escape codes, which will be replaced depending on time and context of the capture call. The string with escapes may be loaded from a template file, using the special syntax (file "path/to/template"). See below for more details.
The rest of the entry is a property list of additional options. Recognized properties are:
Normally new captured information will be appended at the target location (last child, last table line, last list item...). Setting this property will change that.
When set, do not offer to edit the information, just file it away immediately. This makes sense if the template only needs information that can be added automatically.
Set this to the number of lines to insert before and after the new item. Default 0, only common other value is 1.
Start the clock in this item.
Keep the clock running when filing the captured entry.
If starting the capture interrupted a clock, restart that clock when finished with the capture. Note that :clock-keep has precedence over :clock-resume. When setting both to t, the current clock will run and the previous one will not be resumed.
Do not narrow the target buffer, simply show the full buffer. Default is to narrow it so that you only see the new material.
Specification of the location in the table where the new line should be inserted. It should be a string like "II-3" meaning that the new line should become the third line before the second horizontal separator line.
If the target file was not yet visited when capture was invoked, kill the buffer again after capture is completed.

Next: , Previous: Template elements, Up: Capture templates Template expansion

In the template itself, special %-escapes92 allow dynamic insertion of content. The templates are expanded in the order given here:

     %[file]     Insert the contents of the file given by file.
     %(sexp)     Evaluate Elisp sexp and replace with the result.
                       For convenience, %:keyword (see below) placeholders
                       within the expression will be expanded prior to this.
                       The sexp must return a string.
     %<...>      The result of format-time-string on the ... format specification.
     %t          Timestamp, date only.
     %T          Timestamp, with date and time.
     %u, %U      Like the above, but inactive timestamps.
     %i          Initial content, the region when capture is called while the
                 region is active.
                 The entire text will be indented like %i itself.
     %a          Annotation, normally the link created with org-store-link.
     %A          Like %a, but prompt for the description part.
     %l          Like %a, but only insert the literal link.
     %c          Current kill ring head.
     %x          Content of the X clipboard.
     %k          Title of the currently clocked task.
     %K          Link to the currently clocked task.
     %n          User name (taken from user-full-name).
     %f          File visited by current buffer when org-capture was called.
     %F          Full path of the file or directory visited by current buffer.
     %:keyword   Specific information for certain link types, see below.
     %^g         Prompt for tags, with completion on tags in target file.
     %^G         Prompt for tags, with completion all tags in all agenda files.
     %^t         Like %t, but prompt for date.  Similarly %^T, %^u, %^U.
                 You may define a prompt like %^{Birthday}t.
     %^C         Interactive selection of which kill or clip to use.
     %^L         Like %^C, but insert as link.
     %^{prop}p   Prompt the user for a value for property prop.
     %^{prompt}  prompt the user for a string and replace this sequence with it.
                 You may specify a default value and a completion table with
                 The arrow keys access a prompt-specific history.
     %\n         Insert the text entered at the nth %^{prompt}, where n is
                 a number, starting from 1.
     %?          After completing the template, position cursor here.

For specific link types, the following keywords will be defined93:

     Link type                        |  Available keywords
     bbdb                             |  %:name %:company
     irc                              |  %:server %:port %:nick
     vm, vm-imap, wl, mh, mew, rmail  |  %:type %:subject %:message-id
                                      |  %:from %:fromname %:fromaddress
                                      |  %:to   %:toname   %:toaddress
                                      |  %:date (message date header field)
                                      |  %:date-timestamp (date as active timestamp)
                                      |  %:date-timestamp-inactive (date as inactive timestamp)
                                      |  %:fromto (either "to NAME" or "from NAME")94
     gnus                             |  %:group, for messages also all email fields
     w3, w3m                          |  %:url
     info                             |  %:file %:node
     calendar                         |  %:date

To place the cursor after template expansion use:

     %?          After completing the template, position cursor here.

Previous: Template expansion, Up: Capture templates Templates in contexts

To control whether a capture template should be accessible from a specific context, you can customize org-capture-templates-contexts. Let's say for example that you have a capture template "p" for storing Gnus emails containing patches. Then you would configure this option like this:

     (setq org-capture-templates-contexts
           '(("p" (in-mode . "message-mode"))))

You can also tell that the command key "p" should refer to another template. In that case, add this command key like this:

     (setq org-capture-templates-contexts
           '(("p" "q" (in-mode . "message-mode"))))

See the docstring of the variable for more information.

Next: , Previous: Capture, Up: Capture - Refile - Archive

9.2 Attachments

It is often useful to associate reference material with an outline node/task. Small chunks of plain text can simply be stored in the subtree of a project. Hyperlinks (see Hyperlinks) can establish associations with files that live elsewhere on your computer or in the cloud, like emails or source code files belonging to a project. Another method is attachments, which are files located in a directory belonging to an outline node. Org uses directories named by the unique ID of each entry. These directories are located in the data directory which lives in the same directory where your Org file lives95. If you initialize this directory with git init, Org will automatically commit changes when it sees them. The attachment system has been contributed to Org by John Wiegley.

In cases where it seems better to do so, you can also attach a directory of your choice to an entry. You can also make children inherit the attachment directory from a parent, so that an entire subtree uses the same attached directory.

The following commands deal with attachments:

C-c C-a     (org-attach)
The dispatcher for commands related to the attachment system. After these keys, a list of commands is displayed and you must press an additional key to select a command:
a     (org-attach-attach)
Select a file and move it into the task's attachment directory. The file will be copied, moved, or linked, depending on org-attach-method. Note that hard links are not supported on all systems.

Attach a file using the copy/move/link method. Note that hard links are not supported on all systems.

n     (org-attach-new)
Create a new attachment as an Emacs buffer.

z     (org-attach-sync)
Synchronize the current task with its attachment directory, in case you added attachments yourself.

o     (org-attach-open)
Open current task's attachment. If there is more than one, prompt for a file name first. Opening will follow the rules set by org-file-apps. For more details, see the information on following hyperlinks (see Handling links).

O     (org-attach-open-in-emacs)
Also open the attachment, but force opening the file in Emacs.

f     (org-attach-reveal)
Open the current task's attachment directory.

F     (org-attach-reveal-in-emacs)
Also open the directory, but force using dired in Emacs.

d     (org-attach-delete-one)
Select and delete a single attachment.

D     (org-attach-delete-all)
Delete all of a task's attachments. A safer way is to open the directory in dired and delete from there.

s     (org-attach-set-directory)
Set a specific directory as the entry's attachment directory. This works by putting the directory path into the ATTACH_DIR property.

i     (org-attach-set-inherit)
Set the ATTACH_DIR_INHERIT property, so that children will use the same directory for attachments as the parent does.

Next: , Previous: Attachments, Up: Capture - Refile - Archive

9.3 RSS feeds

Org can add and change entries based on information found in RSS feeds and Atom feeds. You could use this to make a task out of each new podcast in a podcast feed. Or you could use a phone-based note-creating service on the web to import tasks into Org. To access feeds, configure the variable org-feed-alist. The docstring of this variable has detailed information. Here is just an example:

     (setq org-feed-alist
              "~/txt/org/" "Slashdot Entries")))

will configure that new items from the feed provided by will result in new entries in the file ~/org/ under the heading ‘Slashdot Entries’, whenever the following command is used:

C-c C-x g     (org-feed-update-all)
C-c C-x g
Collect items from the feeds configured in org-feed-alist and act upon them.
C-c C-x G     (org-feed-goto-inbox)
Prompt for a feed name and go to the inbox configured for this feed.

Under the same headline, Org will create a drawer ‘FEEDSTATUS’ in which it will store information about the status of items in the feed, to avoid adding the same item several times. You should add ‘FEEDSTATUS’ to the list of drawers in that file:


For more information, including how to read atom feeds, see org-feed.el and the docstring of org-feed-alist.

Next: , Previous: RSS Feeds, Up: Capture - Refile - Archive

9.4 Protocols for external access

You can set up Org for handling protocol calls from outside applications that are passed to Emacs through the emacsserver. For example, you can configure bookmarks in your web browser to send a link to the current page to Org and create a note from it using capture (see Capture). Or you could create a bookmark that will tell Emacs to open the local source file of a remote website you are looking at with the browser. See for detailed documentation and setup instructions.

Next: , Previous: Protocols, Up: Capture - Refile - Archive

9.5 Refile and copy

When reviewing the captured data, you may want to refile or to copy some of the entries into a different list, for example into a project. Cutting, finding the right location, and then pasting the note is cumbersome. To simplify this process, you can use the following special command:

C-c M-w     (org-copy)
Copying works like refiling, except that the original note is not deleted.
C-c C-w     (org-refile)
Refile the entry or region at point. This command offers possible locations for refiling the entry and lets you select one with completion. The item (or all items in the region) is filed below the target heading as a subitem. Depending on org-reverse-note-order, it will be either the first or last subitem.
By default, all level 1 headlines in the current buffer are considered to be targets, but you can have more complex definitions across a number of files. See the variable org-refile-targets for details. If you would like to select a location via a file-path-like completion along the outline path, see the variables org-refile-use-outline-path and org-outline-path-complete-in-steps. If you would like to be able to create new nodes as new parents for refiling on the fly, check the variable org-refile-allow-creating-parent-nodes. When the variable org-log-refile96 is set, a timestamp or a note will be recorded when an entry has been refiled.
C-u C-c C-w
Use the refile interface to jump to a heading.
C-u C-u C-c C-w     (org-refile-goto-last-stored)
Jump to the location where org-refile last moved a tree to.
C-2 C-c C-w
Refile as the child of the item currently being clocked.
C-3 C-c C-w
Refile and keep the entry in place. Also see org-refile-keep to make this the default behavior, and beware that this may result in duplicated ID properties.
C-0 C-c C-w or C-u C-u C-u C-c C-w     (org-refile-cache-clear)
Clear the target cache. Caching of refile targets can be turned on by setting org-refile-use-cache. To make the command see new possible targets, you have to clear the cache with this command.

Previous: Refile and copy, Up: Capture - Refile - Archive

9.6 Archiving

When a project represented by a (sub)tree is finished, you may want to move the tree out of the way and to stop it from contributing to the agenda. Archiving is important to keep your working files compact and global searches like the construction of agenda views fast.

C-c C-x C-a     (org-archive-subtree-default)
Archive the current entry using the command specified in the variable org-archive-default-command.

Next: , Previous: Archiving, Up: Archiving

9.6.1 Moving a tree to the archive file

The most common archiving action is to move a project tree to another file, the archive file.

C-c C-x C-s or short C-c $     (org-archive-subtree)
Archive the subtree starting at the cursor position to the location given by org-archive-location.
C-u C-c C-x C-s
Check if any direct children of the current headline could be moved to the archive. To do this, each subtree is checked for open TODO entries. If none are found, the command offers to move it to the archive location. If the cursor is not on a headline when this command is invoked, the level 1 trees will be checked.

The default archive location is a file in the same directory as the current file, with the name derived by appending _archive to the current file name. You can also choose what heading to file archived items under, with the possibility to add them to a datetree in a file. For information and examples on how to specify the file and the heading, see the documentation string of the variable org-archive-location.

There is also an in-buffer option for setting this variable, for example97:

     #+ARCHIVE: %s_done::

If you would like to have a special ARCHIVE location for a single entry or a (sub)tree, give the entry an :ARCHIVE: property with the location as the value (see Properties and Columns).

When a subtree is moved, it receives a number of special properties that record context information like the file from where the entry came, its outline path the archiving time etc. Configure the variable org-archive-save-context-info to adjust the amount of information added.

Previous: Moving subtrees, Up: Archiving

9.6.2 Internal archiving

If you want to just switch off (for agenda views) certain subtrees without moving them to a different file, you can use the ARCHIVE tag.

A headline that is marked with the ARCHIVE tag (see Tags) stays at its location in the outline tree, but behaves in the following way:

The following commands help manage the ARCHIVE tag:

C-c C-x a     (org-toggle-archive-tag)
Toggle the ARCHIVE tag for the current headline. When the tag is set, the headline changes to a shadowed face, and the subtree below it is hidden.
C-u C-c C-x a
Check if any direct children of the current headline should be archived. To do this, each subtree is checked for open TODO entries. If none are found, the command offers to set the ARCHIVE tag for the child. If the cursor is not on a headline when this command is invoked, the level 1 trees will be checked.
C-TAB     (org-force-cycle-archived)
Cycle a tree even if it is tagged with ARCHIVE.
C-c C-x A     (org-archive-to-archive-sibling)
Move the current entry to the Archive Sibling. This is a sibling of the entry with the heading ‘Archive’ and the tag ‘ARCHIVE’. The entry becomes a child of that sibling and in this way retains a lot of its original context, including inherited tags and approximate position in the outline.

Next: , Previous: Capture - Refile - Archive, Up: Top

10 Agenda views

Due to the way Org works, TODO items, time-stamped items, and tagged headlines can be scattered throughout a file or even a number of files. To get an overview of open action items, or of events that are important for a particular date, this information must be collected, sorted and displayed in an organized way.

Org can select items based on various criteria and display them in a separate buffer. Seven different view types are provided:

The extracted information is displayed in a special agenda buffer. This buffer is read-only, but provides commands to visit the corresponding locations in the original Org files, and even to edit these files remotely.

Two variables control how the agenda buffer is displayed and whether the window configuration is restored when the agenda exits: org-agenda-window-setup and org-agenda-restore-windows-after-quit.

Next: , Previous: Agenda Views, Up: Agenda Views

10.1 Agenda files

The information to be shown is normally collected from all agenda files, the files listed in the variable org-agenda-files98. If a directory is part of this list, all files with the extension .org in this directory will be part of the list.

Thus, even if you only work with a single Org file, that file should be put into the list99. You can customize org-agenda-files, but the easiest way to maintain it is through the following commands

C-c [     (org-agenda-file-to-front)
Add current file to the list of agenda files. The file is added to the front of the list. If it was already in the list, it is moved to the front. With a prefix argument, file is added/moved to the end.
C-c ]     (org-remove-file)
Remove current file from the list of agenda files.
C-'     (org-cycle-agenda-files)
Cycle through agenda file list, visiting one file after the other.
M-x org-iswitchb RET
Command to use an iswitchb-like interface to switch to and between Org buffers.

The Org menu contains the current list of files and can be used to visit any of them.

If you would like to focus the agenda temporarily on a file not in this list, or on just one file in the list, or even on only a subtree in a file, then this can be done in different ways. For a single agenda command, you may press < once or several times in the dispatcher (see Agenda dispatcher). To restrict the agenda scope for an extended period, use the following commands:

C-c C-x <     (org-agenda-set-restriction-lock)
Permanently restrict the agenda to the current subtree. When with a prefix argument, or with the cursor before the first headline in a file, the agenda scope is set to the entire file. This restriction remains in effect until removed with C-c C-x >, or by typing either < or > in the agenda dispatcher. If there is a window displaying an agenda view, the new restriction takes effect immediately.
C-c C-x >     (org-agenda-remove-restriction-lock)
Remove the permanent restriction created by C-c C-x <.

When working with speedbar.el, you can use the following commands in the Speedbar frame:

< in the speedbar frame     (org-speedbar-set-agenda-restriction)
Permanently restrict the agenda to the item—either an Org file or a subtree in such a file—at the cursor in the Speedbar frame. If there is a window displaying an agenda view, the new restriction takes effect immediately.
> in the speedbar frame     (org-agenda-remove-restriction-lock)
Lift the restriction.

Next: , Previous: Agenda files, Up: Agenda Views

10.2 The agenda dispatcher

The views are created through a dispatcher, which should be bound to a global key—for example C-c a (see Activation). In the following we will assume that C-c a is indeed how the dispatcher is accessed and list keyboard access to commands accordingly. After pressing C-c a, an additional letter is required to execute a command. The dispatcher offers the following default commands:

Create the calendar-like agenda (see Weekly/daily agenda).
t / T
Create a list of all TODO items (see Global TODO list).
m / M
Create a list of headlines matching a TAGS expression (see Matching tags and properties).
Create the timeline view for the current buffer (see Timeline).
Create a list of entries selected by a boolean expression of keywords and/or regular expressions that must or must not occur in the entry.
Search for a regular expression in all agenda files and additionally in the files listed in org-agenda-text-search-extra-files. This uses the Emacs command multi-occur. A prefix argument can be used to specify the number of context lines for each match, default is 1.
# / !
Create a list of stuck projects (see Stuck projects).
Restrict an agenda command to the current buffer100. After pressing <, you still need to press the character selecting the command.
< <
If there is an active region, restrict the following agenda command to the region. Otherwise, restrict it to the current subtree101. After pressing < <, you still need to press the character selecting the command.
Toggle sticky agenda views. By default, Org maintains only a single agenda buffer and rebuilds it each time you change the view, to make sure everything is always up to date. If you switch between views often and the build time bothers you, you can turn on sticky agenda buffers (make this the default by customizing the variable org-agenda-sticky). With sticky agendas, the dispatcher only switches to the selected view, you need to update it by hand with r or g. You can toggle sticky agenda view any time with org-toggle-sticky-agenda.

You can also define custom commands that will be accessible through the dispatcher, just like the default commands. This includes the possibility to create extended agenda buffers that contain several blocks together, for example the weekly agenda, the global TODO list and a number of special tags matches. See Custom agenda views.

Next: , Previous: Agenda dispatcher, Up: Agenda Views

10.3 The built-in agenda views

In this section we describe the built-in views.

Next: , Previous: Built-in agenda views, Up: Built-in agenda views

10.3.1 The weekly/daily agenda

The purpose of the weekly/daily agenda is to act like a page of a paper agenda, showing all the tasks for the current week or day.

C-c a a     (org-agenda-list)
Compile an agenda for the current week from a list of Org files. The agenda shows the entries for each day. With a numeric prefix102 (like C-u 2 1 C-c a a) you may set the number of days to be displayed.

The default number of days displayed in the agenda is set by the variable org-agenda-span (or the obsolete org-agenda-ndays). This variable can be set to any number of days you want to see by default in the agenda, or to a span name, such as day, week, month or year. For weekly agendas, the default is to start on the previous monday (see org-agenda-start-on-weekday). You can also set the start date using a date shift: (setq org-agenda-start-day "+10d") will start the agenda ten days from today in the future.

Remote editing from the agenda buffer means, for example, that you can change the dates of deadlines and appointments from the agenda buffer. The commands available in the Agenda buffer are listed in Agenda commands.

Calendar/Diary integration

Emacs contains the calendar and diary by Edward M. Reingold. The calendar displays a three-month calendar with holidays from different countries and cultures. The diary allows you to keep track of anniversaries, lunar phases, sunrise/set, recurrent appointments (weekly, monthly) and more. In this way, it is quite complementary to Org. It can be very useful to combine output from Org with the diary.

In order to include entries from the Emacs diary into Org mode's agenda, you only need to customize the variable

     (setq org-agenda-include-diary t)

After that, everything will happen automatically. All diary entries including holidays, anniversaries, etc., will be included in the agenda buffer created by Org mode. <SPC>, <TAB>, and <RET> can be used from the agenda buffer to jump to the diary file in order to edit existing diary entries. The i command to insert new entries for the current date works in the agenda buffer, as well as the commands S, M, and C to display Sunrise/Sunset times, show lunar phases and to convert to other calendars, respectively. c can be used to switch back and forth between calendar and agenda.

If you are using the diary only for sexp entries and holidays, it is faster to not use the above setting, but instead to copy or even move the entries into an Org file. Org mode evaluates diary-style sexp entries, and does it faster because there is no overhead for first creating the diary display. Note that the sexp entries must start at the left margin, no whitespace is allowed before them. For example, the following segment of an Org file will be processed and entries will be made in the agenda:

     * Birthdays and similar stuff
     #+CATEGORY: Holiday
     %%(org-calendar-holiday)   ; special function for holiday names
     #+CATEGORY: Ann
     %%(org-anniversary 1956  5 14)103 Arthur Dent is %d years old
     %%(org-anniversary 1869 10  2) Mahatma Gandhi would be %d years old
Anniversaries from BBDB

If you are using the Big Brothers Database to store your contacts, you will very likely prefer to store anniversaries in BBDB rather than in a separate Org or diary file. Org supports this and will show BBDB anniversaries as part of the agenda. All you need to do is to add the following to one of your agenda files:

     * Anniversaries
       :CATEGORY: Anniv

You can then go ahead and define anniversaries for a BBDB record. Basically, you need to press C-o anniversary <RET> with the cursor in a BBDB record and then add the date in the format YYYY-MM-DD or MM-DD, followed by a space and the class of the anniversary (‘birthday’ or ‘wedding’, or a format string). If you omit the class, it will default to ‘birthday’. Here are a few examples, the header for the file org-bbdb.el contains more detailed information.

     1955-08-02 wedding
     2008-04-14 %s released version 6.01 of org mode, %d years ago

After a change to BBDB, or for the first agenda display during an Emacs session, the agenda display will suffer a short delay as Org updates its hash with anniversaries. However, from then on things will be very fast—much faster in fact than a long list of ‘%%(diary-anniversary)’ entries in an Org or Diary file.

Appointment reminders

Org can interact with Emacs appointments notification facility. To add the appointments of your agenda files, use the command org-agenda-to-appt. This command lets you filter through the list of your appointments and add only those belonging to a specific category or matching a regular expression. It also reads a APPT_WARNTIME property which will then override the value of appt-message-warning-time for this appointment. See the docstring for details.

Next: , Previous: Weekly/daily agenda, Up: Built-in agenda views

10.3.2 The global TODO list

The global TODO list contains all unfinished TODO items formatted and collected into a single place.

C-c a t     (org-todo-list)
Show the global TODO list. This collects the TODO items from all agenda files (see Agenda Views) into a single buffer. By default, this lists items with a state the is not a DONE state. The buffer is in agenda-mode, so there are commands to examine and manipulate the TODO entries directly from that buffer (see Agenda commands).
C-c a T     (org-todo-list)
Like the above, but allows selection of a specific TODO keyword. You can also do this by specifying a prefix argument to C-c a t. You are prompted for a keyword, and you may also specify several keywords by separating them with ‘|’ as the boolean OR operator. With a numeric prefix, the Nth keyword in org-todo-keywords is selected. The r key in the agenda buffer regenerates it, and you can give a prefix argument to this command to change the selected TODO keyword, for example 3 r. If you often need a search for a specific keyword, define a custom command for it (see Agenda dispatcher).
Matching specific TODO keywords can also be done as part of a tags search (see Tag searches).

Remote editing of TODO items means that you can change the state of a TODO entry with a single key press. The commands available in the TODO list are described in Agenda commands.

Normally the global TODO list simply shows all headlines with TODO keywords. This list can become very long. There are two ways to keep it more compact:

Next: , Previous: Global TODO list, Up: Built-in agenda views

10.3.3 Matching tags and properties

If headlines in the agenda files are marked with tags (see Tags), or have properties (see Properties and Columns), you can select headlines based on this metadata and collect them into an agenda buffer. The match syntax described here also applies when creating sparse trees with C-c / m.

C-c a m     (org-tags-view)
Produce a list of all headlines that match a given set of tags. The command prompts for a selection criterion, which is a boolean logic expression with tags, like ‘+work+urgent-withboss’ or ‘work|home’ (see Tags). If you often need a specific search, define a custom command for it (see Agenda dispatcher).
C-c a M     (org-tags-view)
Like C-c a m, but only select headlines that are also TODO items in a not-DONE state and force checking subitems (see variable org-tags-match-list-sublevels). To exclude scheduled/deadline items, see the variable org-agenda-tags-todo-honor-ignore-options. Matching specific TODO keywords together with a tags match is also possible, see Tag searches.

The commands available in the tags list are described in Agenda commands.

Match syntax

A search string can use Boolean operators ‘&’ for AND and ‘|’ for OR. ‘&’ binds more strongly than ‘|’. Parentheses are not implemented. Each element in the search is either a tag, a regular expression matching tags, or an expression like PROPERTY OPERATOR VALUE with a comparison operator, accessing a property value. Each element may be preceded by ‘-’, to select against it, and ‘+’ is syntactic sugar for positive selection. The AND operator ‘&’ is optional when ‘+’ or ‘-’ is present. Here are some examples, using only tags.

Select headlines tagged ‘:work:’.
Select headlines tagged ‘:work:’ and ‘:boss:’.
Select headlines tagged ‘:work:’, but discard those also tagged ‘:boss:’.
Selects lines tagged ‘:work:’ or ‘:laptop:’.
Like before, but require the ‘:laptop:’ lines to be tagged also ‘:night:’.

Instead of a tag, you may also specify a regular expression enclosed in curly braces. For example, ‘work+{^boss.*}’ matches headlines that contain the tag ‘:work:’ and any tag starting with ‘boss’.

Group tags (see Tag groups) are expanded as regular expressions. E.g., if ‘:work:’ is a group tag for the group ‘:work:lab:conf:’, then searching for ‘work’ will search for ‘{\(?:work\|lab\|conf\)}’ and searching for ‘-work’ will search for all headlines but those with one of the tag in the group (i.e., ‘-{\(?:work\|lab\|conf\)}’).

You may also test for properties (see Properties and Columns) at the same time as matching tags. The properties may be real properties, or special properties that represent other metadata (see Special properties). For example, the “property” TODO represents the TODO keyword of the entry and the “property” PRIORITY represents the PRIORITY keyword of the entry. The ITEM special property cannot currently be used in tags/property searches104.

Except the see Special properties, one other “property” can also be used. LEVEL represents the level of an entry. So a search ‘+LEVEL=3+boss-TODO="DONE"’ lists all level three headlines that have the tag ‘boss’ and are not marked with the TODO keyword DONE. In buffers with org-odd-levels-only set, ‘LEVEL’ does not count the number of stars, but ‘LEVEL=2’ will correspond to 3 stars etc.

Here are more examples:

Select ‘:work:’-tagged TODO lines with the specific TODO keyword ‘WAITING’.
Waiting tasks both at work and at home.

When matching properties, a number of different operators can be used to test the value of a property. Here is a complex example:

     +work-boss+PRIORITY="A"+Coffee="unlimited"+Effort<2         \

The type of comparison will depend on how the comparison value is written:

So the search string in the example finds entries tagged ‘:work:’ but not ‘:boss:’, which also have a priority value ‘A’, a ‘:Coffee:’ property with the value ‘unlimited’, an ‘Effort’ property that is numerically smaller than 2, a ‘:With:’ property that is matched by the regular expression ‘Sarah\|Denny’, and that are scheduled on or after October 11, 2008.

Accessing TODO, LEVEL, and CATEGORY during a search is fast. Accessing any other properties will slow down the search. However, once you have paid the price by accessing one property, testing additional properties is cheap again.

You can configure Org mode to use property inheritance during a search, but beware that this can slow down searches considerably. See Property inheritance, for details.

For backward compatibility, and also for typing speed, there is also a different way to test TODO states in a search. For this, terminate the tags/property part of the search string (which may include several terms connected with ‘|’) with a ‘/’ and then specify a Boolean expression just for TODO keywords. The syntax is then similar to that for tags, but should be applied with care: for example, a positive selection on several TODO keywords cannot meaningfully be combined with boolean AND. However, negative selection combined with AND can be meaningful. To make sure that only lines are checked that actually have any TODO keyword (resulting in a speed-up), use C-c a M, or equivalently start the TODO part after the slash with ‘!’. Using C-c a M or ‘/!’ will not match TODO keywords in a DONE state. Examples:

Same as ‘work+TODO="WAITING"
Select ‘:work:’-tagged TODO lines that are neither ‘WAITING’ nor ‘NEXT
Select ‘:work:’-tagged TODO lines that are either ‘WAITING’ or ‘NEXT’.

Next: , Previous: Matching tags and properties, Up: Built-in agenda views

10.3.4 Timeline for a single file

The timeline summarizes all time-stamped items from a single Org mode file in a time-sorted view. The main purpose of this command is to give an overview over events in a project.

C-c a L     (org-timeline)
Show a time-sorted view of the Org file, with all time-stamped items. When called with a C-u prefix, all unfinished TODO entries (scheduled or not) are also listed under the current date.

The commands available in the timeline buffer are listed in Agenda commands.

Next: , Previous: Timeline, Up: Built-in agenda views

10.3.5 Search view

This agenda view is a general text search facility for Org mode entries. It is particularly useful to find notes.

C-c a s     (org-search-view)
This is a special search that lets you select entries by matching a substring or specific words using a boolean logic.
For example, the search string ‘computer equipment’ will find entries that contain ‘computer equipment’ as a substring. If the two words are separated by more space or a line break, the search will still match. Search view can also search for specific keywords in the entry, using Boolean logic. The search string ‘+computer +wifi -ethernet -{8\.11[bg]}’ will search for note entries that contain the keywords computer and wifi, but not the keyword ethernet, and which are also not matched by the regular expression 8\.11[bg], meaning to exclude both 8.11b and 8.11g. The first ‘+’ is necessary to turn on word search, other ‘+’ characters are optional. For more details, see the docstring of the command org-search-view.

Note that in addition to the agenda files, this command will also search the files listed in org-agenda-text-search-extra-files.

Previous: Search view, Up: Built-in agenda views

10.3.6 Stuck projects

If you are following a system like David Allen's GTD to organize your work, one of the “duties” you have is a regular review to make sure that all projects move along. A stuck project is a project that has no defined next actions, so it will never show up in the TODO lists Org mode produces. During the review, you need to identify such projects and define next actions for them.

C-c a #     (org-agenda-list-stuck-projects)
List projects that are stuck.
C-c a !
Customize the variable org-stuck-projects to define what a stuck project is and how to find it.

You almost certainly will have to configure this view before it will work for you. The built-in default assumes that all your projects are level-2 headlines, and that a project is not stuck if it has at least one entry marked with a TODO keyword TODO or NEXT or NEXTACTION.

Let's assume that you, in your own way of using Org mode, identify projects with a tag PROJECT, and that you use a TODO keyword MAYBE to indicate a project that should not be considered yet. Let's further assume that the TODO keyword DONE marks finished projects, and that NEXT and TODO indicate next actions. The tag @SHOP indicates shopping and is a next action even without the NEXT tag. Finally, if the project contains the special word IGNORE anywhere, it should not be listed either. In this case you would start by identifying eligible projects with a tags/todo match105+PROJECT/-MAYBE-DONE’, and then check for TODO, NEXT, @SHOP, and IGNORE in the subtree to identify projects that are not stuck. The correct customization for this is

     (setq org-stuck-projects
           '("+PROJECT/-MAYBE-DONE" ("NEXT" "TODO") ("@SHOP")

Note that if a project is identified as non-stuck, the subtree of this entry will still be searched for stuck projects.

Next: , Previous: Built-in agenda views, Up: Agenda Views

10.4 Presentation and sorting

Before displaying items in an agenda view, Org mode visually prepares the items and sorts them. Each item occupies a single line. The line starts with a prefix that contains the category (see Categories) of the item and other important information. You can customize in which column tags will be displayed through org-agenda-tags-column. You can also customize the prefix using the option org-agenda-prefix-format. This prefix is followed by a cleaned-up version of the outline headline associated with the item.

Next: , Previous: Presentation and sorting, Up: Presentation and sorting

10.4.1 Categories

The category is a broad label assigned to each agenda item. By default, the category is simply derived from the file name, but you can also specify it with a special line in the buffer, like this106:

     #+CATEGORY: Thesis

If you would like to have a special CATEGORY for a single entry or a (sub)tree, give the entry a :CATEGORY: property with the special category you want to apply as the value.

The display in the agenda buffer looks best if the category is not longer than 10 characters.

You can set up icons for category by customizing the org-agenda-category-icon-alist variable.

Next: , Previous: Categories, Up: Presentation and sorting

10.4.2 Time-of-day specifications

Org mode checks each agenda item for a time-of-day specification. The time can be part of the timestamp that triggered inclusion into the agenda, for example as in ‘<2005-05-10 Tue 19:00>. Time ranges can be specified with two timestamps, like ‘<2005-05-10 Tue 20:30>--<2005-05-10 Tue 22:15>.

In the headline of the entry itself, a time(range) may also appear as plain text (like ‘12:45’ or a ‘8:30-1pm’). If the agenda integrates the Emacs diary (see Weekly/daily agenda), time specifications in diary entries are recognized as well.

For agenda display, Org mode extracts the time and displays it in a standard 24 hour format as part of the prefix. The example times in the previous paragraphs would end up in the agenda like this:

         8:30-13:00 Arthur Dent lies in front of the bulldozer
        12:45...... Ford Prefect arrives and takes Arthur to the pub
        19:00...... The Vogon reads his poem
        20:30-22:15 Marvin escorts the Hitchhikers to the bridge

If the agenda is in single-day mode, or for the display of today, the timed entries are embedded in a time grid, like

         8:00...... ------------------
         8:30-13:00 Arthur Dent lies in front of the bulldozer
        10:00...... ------------------
        12:00...... ------------------
        12:45...... Ford Prefect arrives and takes Arthur to the pub
        14:00...... ------------------
        16:00...... ------------------
        18:00...... ------------------
        19:00...... The Vogon reads his poem
        20:00...... ------------------
        20:30-22:15 Marvin escorts the Hitchhikers to the bridge

The time grid can be turned on and off with the variable org-agenda-use-time-grid, and can be configured with org-agenda-time-grid.

Next: , Previous: Time-of-day specifications, Up: Presentation and sorting

10.4.3 Sorting agenda items

Before being inserted into a view, the items are sorted. How this is done depends on the type of view.

Sorting can be customized using the variable org-agenda-sorting-strategy, and may also include criteria based on the estimated effort of an entry (see Effort estimates).

Previous: Sorting agenda items, Up: Presentation and sorting

10.4.4 Filtering/limiting agenda items

Agenda built-in or customized commands are statically defined. Agenda filters and limits provide two ways of dynamically narrowing down the list of agenda entries: fitlers and limits. Filters only act on the display of the items, while limits take effect before the list of agenda entries is built. Filter are more often used interactively, while limits are mostly useful when defined as local variables within custom agenda commands.

Filtering in the agenda

/     (org-agenda-filter-by-tag)
Filter the agenda view with respect to a tag and/or effort estimates. The difference between this and a custom agenda command is that filtering is very fast, so that you can switch quickly between different filters without having to recreate the agenda.107

You will be prompted for a tag selection letter; <SPC> will mean any tag at all. Pressing <TAB> at that prompt will offer use completion to select a tag (including any tags that do not have a selection character). The command then hides all entries that do not contain or inherit this tag. When called with prefix arg, remove the entries that do have the tag. A second / at the prompt will turn off the filter and unhide any hidden entries. If the first key you press is either + or -, the previous filter will be narrowed by requiring or forbidding the selected additional tag. Instead of pressing + or - after /, you can also immediately use the \ command.

In order to filter for effort estimates, you should set up allowed efforts globally, for example

          (setq org-global-properties
              '(("Effort_ALL". "0 0:10 0:30 1:00 2:00 3:00 4:00")))

You can then filter for an effort by first typing an operator, one of <, >, and =, and then the one-digit index of an effort estimate in your array of allowed values, where 0 means the 10th value. The filter will then restrict to entries with effort smaller-or-equal, equal, or larger-or-equal than the selected value. If the digits 0–9 are not used as fast access keys to tags, you can also simply press the index digit directly without an operator. In this case, < will be assumed. For application of the operator, entries without a defined effort will be treated according to the value of org-sort-agenda-noeffort-is-high. To filter for tasks without effort definition, press ? as the operator.

Org also supports automatic, context-aware tag filtering. If the variable org-agenda-auto-exclude-function is set to a user-defined function, that function can decide which tags should be excluded from the agenda automatically. Once this is set, the / command then accepts RET as a sub-option key and runs the auto exclusion logic. For example, let's say you use a Net tag to identify tasks which need network access, an Errand tag for errands in town, and a Call tag for making phone calls. You could auto-exclude these tags based on the availability of the Internet, and outside of business hours, with something like this:

          (defun org-my-auto-exclude-function (tag)
            (and (cond
                  ((string= tag "Net")
                   (/= 0 (call-process "/sbin/ping" nil nil nil
                                       "-c1" "-q" "-t1" "")))
                  ((or (string= tag "Errand") (string= tag "Call"))
                   (let ((hour (nth 2 (decode-time))))
                     (or (< hour 8) (> hour 21)))))
                 (concat "-" tag)))
          (setq org-agenda-auto-exclude-function 'org-my-auto-exclude-function)

\     (org-agenda-filter-by-tag-refine)
Narrow the current agenda filter by an additional condition. When called with prefix arg, remove the entries that do have the tag, or that do match the effort criterion. You can achieve the same effect by pressing + or - as the first key after the / command.

[ ] { }
in search view
add new search words ([ and ]) or new regular expressions ({ and }) to the query string. The opening bracket/brace will add a positive search term prefixed by ‘+’, indicating that this search term must occur/match in the entry. The closing bracket/brace will add a negative search term which must not occur/match in the entry for it to be selected.

<     (org-agenda-filter-by-category)
Filter the current agenda view with respect to the category of the item at point. Pressing < another time will remove this filter. You can add a filter preset through the option org-agenda-category-filter-preset (see below.)

^     (org-agenda-filter-by-top-headline)
Filter the current agenda view and only display the siblings and the parent headline of the one at point.

=     (org-agenda-filter-by-regexp)
Filter the agenda view by a regular expression: only show agenda entries matching the regular expression the user entered. When called with a prefix argument, it will filter out entries matching the regexp. With two universal prefix arguments, it will remove all the regexp filters, which can be accumulated. You can add a filter preset through the option org-agenda-category-filter-preset (see below.)

|     (org-agenda-filter-remove-all)
Remove all filters in the current agenda view.
Setting limits for the agenda

Here is a list of options that you can set, either globally, or locally in your custom agenda viewssee Custom agenda views.

Limit the number of entries.
Limit the duration of accumulated efforts (as minutes).
Limit the number of entries with TODO keywords.
Limit the number of tagged entries.

When set to a positive integer, each option will exclude entries from other categories: for example, (setq org-agenda-max-effort 100) will limit the agenda to 100 minutes of effort and exclude any entry that as no effort property. If you want to include entries with no effort property, use a negative value for org-agenda-max-effort.

One useful setup is to use org-agenda-max-entries locally in a custom command. For example, this custom command will display the next five entries with a NEXT TODO keyword.

     (setq org-agenda-custom-commands
           '(("n" todo "NEXT"
              ((org-agenda-max-entries 5)))))

Once you mark one of these five entry as DONE, rebuilding the agenda will again the next five entries again, including the first entry that was excluded so far.

You can also dynamically set temporary limits108:

~     (org-agenda-limit-interactively)
This prompts for the type of limit to apply and its value.

Next: , Previous: Presentation and sorting, Up: Agenda Views

10.5 Commands in the agenda buffer

Entries in the agenda buffer are linked back to the Org file or diary file where they originate. You are not allowed to edit the agenda buffer itself, but commands are provided to show and jump to the original entry location, and to edit the Org files “remotely” from the agenda buffer. In this way, all information is stored only once, removing the risk that your agenda and note files may diverge.

Some commands can be executed with mouse clicks on agenda lines. For the other commands, the cursor needs to be in the desired line.


n     (org-agenda-next-line)
Next line (same as <down> and C-n).
p     (org-agenda-previous-line)
Previous line (same as <up> and C-p).
View/Go to Org file

<SPC> or mouse-3     (org-agenda-show-and-scroll-up)
Display the original location of the item in another window. With prefix arg, make sure that the entire entry is made visible in the outline, not only the heading.
L     (org-agenda-recenter)
Display original location and recenter that window.
<TAB> or mouse-2     (org-agenda-goto)
Go to the original location of the item in another window.
<RET>     (org-agenda-switch-to)
Go to the original location of the item and delete other windows.
F     (org-agenda-follow-mode)
Toggle Follow mode. In Follow mode, as you move the cursor through the agenda buffer, the other window always shows the corresponding location in the Org file. The initial setting for this mode in new agenda buffers can be set with the variable org-agenda-start-with-follow-mode.
C-c C-x b     (org-agenda-tree-to-indirect-buffer)
Display the entire subtree of the current item in an indirect buffer. With a numeric prefix argument N, go up to level N and then take that tree. If N is negative, go up that many levels. With a C-u prefix, do not remove the previously used indirect buffer.

C-c C-o     (org-agenda-open-link)
Follow a link in the entry. This will offer a selection of any links in the text belonging to the referenced Org node. If there is only one link, it will be followed without a selection prompt.
Change display

Interactively select another agenda view and append it to the current view.
Delete other windows.
v d or short d     (org-agenda-day-view)
v w or short w     (org-agenda-week-view)
v t     (org-agenda-fortnight-view)
v m     (org-agenda-month-view)
v y     (org-agenda-year-view)
v SPC     (org-agenda-reset-view)
Switch to day/week/month/year view. When switching to day or week view, this setting becomes the default for subsequent agenda refreshes. Since month and year views are slow to create, they do not become the default. A numeric prefix argument may be used to jump directly to a specific day of the year, ISO week, month, or year, respectively. For example, 32 d jumps to February 1st, 9 w to ISO week number 9. When setting day, week, or month view, a year may be encoded in the prefix argument as well. For example, 200712 w will jump to week 12 in 2007. If such a year specification has only one or two digits, it will be mapped to the interval 1938–2037. v <SPC> will reset to what is set in org-agenda-span.
f     (org-agenda-later)
Go forward in time to display the following org-agenda-current-span days. For example, if the display covers a week, switch to the following week. With prefix arg, go forward that many times org-agenda-current-span days.
b     (org-agenda-earlier)
Go backward in time to display earlier dates.
.     (org-agenda-goto-today)
Go to today.
j     (org-agenda-goto-date)
Prompt for a date and go there.
J     (org-agenda-clock-goto)
Go to the currently clocked-in task in the agenda buffer.
D     (org-agenda-toggle-diary)
Toggle the inclusion of diary entries. See Weekly/daily agenda.
v l or short l     (org-agenda-log-mode)
Toggle Logbook mode. In Logbook mode, entries that were marked DONE while logging was on (variable org-log-done) are shown in the agenda, as are entries that have been clocked on that day. You can configure the entry types that should be included in log mode using the variable org-agenda-log-mode-items. When called with a C-u prefix, show all possible logbook entries, including state changes. When called with two prefix arguments C-u C-u, show only logging information, nothing else. v L is equivalent to C-u v l.
v [ or short [     (org-agenda-manipulate-query-add)
Include inactive timestamps into the current view. Only for weekly/daily agenda and timeline views.
v a     (org-agenda-archives-mode)
v A     (org-agenda-archives-mode 'files)
Toggle Archives mode. In Archives mode, trees that are marked ARCHIVED are also scanned when producing the agenda. When you use the capital A, even all archive files are included. To exit archives mode, press v a again.
v R or short R     (org-agenda-clockreport-mode)
Toggle Clockreport mode. In Clockreport mode, the daily/weekly agenda will always show a table with the clocked times for the time span and file scope covered by the current agenda view. The initial setting for this mode in new agenda buffers can be set with the variable org-agenda-start-with-clockreport-mode. By using a prefix argument when toggling this mode (i.e., C-u R), the clock table will not show contributions from entries that are hidden by agenda filtering109. See also the variable org-clock-report-include-clocking-task.
v c
Show overlapping clock entries, clocking gaps, and other clocking problems in the current agenda range. You can then visit clocking lines and fix them manually. See the variable org-agenda-clock-consistency-checks for information on how to customize the definition of what constituted a clocking problem. To return to normal agenda display, press l to exit Logbook mode.
v E or short E     (org-agenda-entry-text-mode)
Toggle entry text mode. In entry text mode, a number of lines from the Org outline node referenced by an agenda line will be displayed below the line. The maximum number of lines is given by the variable org-agenda-entry-text-maxlines. Calling this command with a numeric prefix argument will temporarily modify that number to the prefix value.
G     (org-agenda-toggle-time-grid)
Toggle the time grid on and off. See also the variables org-agenda-use-time-grid and org-agenda-time-grid.
r     (org-agenda-redo)
Recreate the agenda buffer, for example to reflect the changes after modification of the timestamps of items with S-<left> and S-<right>. When the buffer is the global TODO list, a prefix argument is interpreted to create a selective list for a specific TODO keyword.
g     (org-agenda-redo)
Same as r.
C-x C-s or short s     (org-save-all-org-buffers)
Save all Org buffers in the current Emacs session, and also the locations of IDs.
C-c C-x C-c     (org-agenda-columns)
Invoke column view (see Column view) in the agenda buffer. The column view format is taken from the entry at point, or (if there is no entry at point), from the first entry in the agenda view. So whatever the format for that entry would be in the original buffer (taken from a property, from a #+COLUMNS line, or from the default variable org-columns-default-format), will be used in the agenda.

C-c C-x >     (org-agenda-remove-restriction-lock)
Remove the restriction lock on the agenda, if it is currently restricted to a file or subtree (see Agenda files).
Secondary filtering and query editing

For a detailed description of these commands, see see Filtering/limiting agenda items.

/     (org-agenda-filter-by-tag)
Filter the agenda view with respect to a tag and/or effort estimates.

\     (org-agenda-filter-by-tag-refine)
Narrow the current agenda filter by an additional condition.

<     (org-agenda-filter-by-category)
Filter the current agenda view with respect to the category of the item at point. Pressing < another time will remove this filter.

^     (org-agenda-filter-by-top-headline)
Filter the current agenda view and only display the siblings and the parent headline of the one at point.

=     (org-agenda-filter-by-regexp)
Filter the agenda view by a regular expression: only show agenda entries matching the regular expression the user entered. When called with a prefix argument, it will filter out entries matching the regexp. With two universal prefix arguments, it will remove all the regexp filters, which can be accumulated. You can add a filter preset through the option org-agenda-category-filter-preset (see below.)

|     (org-agenda-filter-remove-all)
Remove all filters in the current agenda view.
Remote editing

Digit argument.
C-_     (org-agenda-undo)
Undo a change due to a remote editing command. The change is undone both in the agenda buffer and in the remote buffer.
t     (org-agenda-todo)
Change the TODO state of the item, both in the agenda and in the original org file.
C-S-<right>     (org-agenda-todo-nextset)

C-S-<left>     (org-agenda-todo-previousset)
Switch to the next/previous set of TODO keywords.
C-k     (org-agenda-kill)
Delete the current agenda item along with the entire subtree belonging to it in the original Org file. If the text to be deleted remotely is longer than one line, the kill needs to be confirmed by the user. See variable org-agenda-confirm-kill.
C-c C-w     (org-agenda-refile)
Refile the entry at point.
C-c C-x C-a or short a     (org-agenda-archive-default-with-confirmation)
Archive the subtree corresponding to the entry at point using the default archiving command set in org-archive-default-command. When using the a key, confirmation will be required.
C-c C-x a     (org-agenda-toggle-archive-tag)
Toggle the ARCHIVE tag for the current headline.
C-c C-x A     (org-agenda-archive-to-archive-sibling)
Move the subtree corresponding to the current entry to its archive sibling.
C-c C-x C-s or short $     (org-agenda-archive)
Archive the subtree corresponding to the current headline. This means the entry will be moved to the configured archive location, most likely a different file.
T     (org-agenda-show-tags)
Show all tags associated with the current item. This is useful if you have turned off org-agenda-show-inherited-tags, but still want to see all tags of a headline occasionally.
:     (org-agenda-set-tags)
Set tags for the current headline. If there is an active region in the agenda, change a tag for all headings in the region.
Set the priority for the current item (org-agenda-priority). Org mode prompts for the priority character. If you reply with <SPC>, the priority cookie is removed from the entry.
P     (org-agenda-show-priority)
Display weighted priority of current item.
+ or S-<up>     (org-agenda-priority-up)
Increase the priority of the current item. The priority is changed in the original buffer, but the agenda is not resorted. Use the r key for this.
- or S-<down>     (org-agenda-priority-down)
Decrease the priority of the current item.
z or C-c C-z     (org-agenda-add-note)
Add a note to the entry. This note will be recorded, and then filed to the same location where state change notes are put. Depending on org-log-into-drawer, this may be inside a drawer.
C-c C-a     (org-attach)
Dispatcher for all command related to attachments.
C-c C-s     (org-agenda-schedule)
Schedule this item. With prefix arg remove the scheduling timestamp
C-c C-d     (org-agenda-deadline)
Set a deadline for this item. With prefix arg remove the deadline.
S-<right>     (org-agenda-do-date-later)
Change the timestamp associated with the current line by one day into the future. If the date is in the past, the first call to this command will move it to today.
With a numeric prefix argument, change it by that many days. For example, 3 6 5 S-<right> will change it by a year. With a C-u prefix, change the time by one hour. If you immediately repeat the command, it will continue to change hours even without the prefix arg. With a double C-u C-u prefix, do the same for changing minutes.
The stamp is changed in the original Org file, but the change is not directly reflected in the agenda buffer. Use r or g to update the buffer.
S-<left>     (org-agenda-do-date-earlier)
Change the timestamp associated with the current line by one day into the past.
>     (org-agenda-date-prompt)
Change the timestamp associated with the current line. The key > has been chosen, because it is the same as S-. on my keyboard.
I     (org-agenda-clock-in)
Start the clock on the current item. If a clock is running already, it is stopped first.
O     (org-agenda-clock-out)
Stop the previously started clock.
X     (org-agenda-clock-cancel)
Cancel the currently running clock.
J     (org-agenda-clock-goto)
Jump to the running clock in another window.
k     (org-agenda-capture)
Like org-capture, but use the date at point as the default date for the capture template. See org-capture-use-agenda-date to make this the default behavior of org-capture.
Dragging agenda lines forward/backward

M-<up>     (org-agenda-drag-line-backward)
Drag the line at point backward one line110. With a numeric prefix argument, drag backward by that many lines.

M-<down>     (org-agenda-drag-line-forward)
Drag the line at point forward one line. With a numeric prefix argument, drag forward by that many lines.
Bulk remote editing selected entries

m     (org-agenda-bulk-mark)
Mark the entry at point for bulk action. With numeric prefix argument, mark that many successive entries.
*     (org-agenda-bulk-mark-all)
Mark all visible agenda entries for bulk action.
u     (org-agenda-bulk-unmark)
Unmark entry at point for bulk action.
U     (org-agenda-bulk-remove-all-marks)
Unmark all marked entries for bulk action.
M-m     (org-agenda-bulk-toggle)
Toggle mark of the entry at point for bulk action.
M-*     (org-agenda-bulk-toggle-all)
Toggle marks of all visible entries for bulk action.
%     (org-agenda-bulk-mark-regexp)
Mark entries matching a regular expression for bulk action.
B     (org-agenda-bulk-action)
Bulk action: act on all marked entries in the agenda. This will prompt for another key to select the action to be applied. The prefix arg to B will be passed through to the s and d commands, to bulk-remove these special timestamps. By default, marks are removed after the bulk. If you want them to persist, set org-agenda-bulk-persistent-marks to t or hit p at the prompt.
Toggle persistent marks.
Archive all selected entries.
Archive entries by moving them to their respective archive siblings.
Change TODO state. This prompts for a single TODO keyword and changes the state of all selected entries, bypassing blocking and suppressing logging notes (but not timestamps).
Add a tag to all selected entries.
Remove a tag from all selected entries.
Schedule all items to a new date. To shift existing schedule dates by a fixed number of days, use something starting with double plus at the prompt, for example ‘++8d’ or ‘++2w’.
Set deadline to a specific date.
Prompt for a single refile target and move all entries. The entries will no longer be in the agenda; refresh (g) to bring them back.
Reschedule randomly into the coming N days. N will be prompted for. With prefix arg (C-u B S), scatter only across weekdays.
Apply a function111 to marked entries. For example, the function below sets the CATEGORY property of the entries to web.
               (defun set-category ()
                 (interactive "P")
                 (let* ((marker (or (org-get-at-bol 'org-hd-marker)
                        (buffer (marker-buffer marker)))
                   (with-current-buffer buffer
                         (goto-char marker)
                         (org-back-to-heading t)
                         (org-set-property "CATEGORY" "web"))))))

Calendar commands

c     (org-agenda-goto-calendar)
Open the Emacs calendar and move to the date at the agenda cursor.
c     (org-calendar-goto-agenda)
When in the calendar, compute and show the Org mode agenda for the date at the cursor.
i     (org-agenda-diary-entry)
Insert a new entry into the diary, using the date at the cursor and (for block entries) the date at the mark. This will add to the Emacs diary file112, in a way similar to the i command in the calendar. The diary file will pop up in another window, where you can add the entry.

If you configure org-agenda-diary-file to point to an Org mode file, Org will create entries (in Org mode syntax) in that file instead. Most entries will be stored in a date-based outline tree that will later make it easy to archive appointments from previous months/years. The tree will be built under an entry with a DATE_TREE property, or else with years as top-level entries. Emacs will prompt you for the entry text—if you specify it, the entry will be created in org-agenda-diary-file without further interaction. If you directly press <RET> at the prompt without typing text, the target file will be shown in another window for you to finish the entry there. See also the k r command.

M     (org-agenda-phases-of-moon)
Show the phases of the moon for the three months around current date.
S     (org-agenda-sunrise-sunset)
Show sunrise and sunset times. The geographical location must be set with calendar variables, see the documentation for the Emacs calendar.
C     (org-agenda-convert-date)
Convert the date at cursor into many other cultural and historic calendars.
H     (org-agenda-holidays)
Show holidays for three months around the cursor date.
M-x org-icalendar-combine-agenda-files RET
Export a single iCalendar file containing entries from all agenda files. This is a globally available command, and also available in the agenda menu.
Exporting to a file

C-x C-w     (org-agenda-write)
Write the agenda view to a file. Depending on the extension of the selected file name, the view will be exported as HTML (.html or .htm), Postscript (.ps), PDF (.pdf), Org (.org) and plain text (any other extension). When exporting to Org, only the body of original headlines are exported, not subtrees or inherited tags. When called with a C-u prefix argument, immediately open the newly created file. Use the variable org-agenda-exporter-settings to set options for ps-print and for htmlize to be used during export.
Quit and Exit

q     (org-agenda-quit)
Quit agenda, remove the agenda buffer.
x     (org-agenda-exit)
Exit agenda, remove the agenda buffer and all buffers loaded by Emacs for the compilation of the agenda. Buffers created by the user to visit Org files will not be removed.

Next: , Previous: Agenda commands, Up: Agenda Views

10.6 Custom agenda views

Custom agenda commands serve two purposes: to store and quickly access frequently used TODO and tags searches, and to create special composite agenda buffers. Custom agenda commands will be accessible through the dispatcher (see Agenda dispatcher), just like the default commands.

Next: , Previous: Custom agenda views, Up: Custom agenda views

10.6.1 Storing searches

The first application of custom searches is the definition of keyboard shortcuts for frequently used searches, either creating an agenda buffer, or a sparse tree (the latter covering of course only the current buffer). Custom commands are configured in the variable org-agenda-custom-commands. You can customize this variable, for example by pressing C-c a C. You can also directly set it with Emacs Lisp in .emacs. The following example contains all valid agenda views:

     (setq org-agenda-custom-commands
           '(("x" agenda)
             ("y" agenda*)
             ("w" todo "WAITING")
             ("W" todo-tree "WAITING")
             ("u" tags "+boss-urgent")
             ("v" tags-todo "+boss-urgent")
             ("U" tags-tree "+boss-urgent")
             ("f" occur-tree "\\<FIXME\\>")
             ("h" . "HOME+Name tags searches") ; description for "h" prefix
             ("hl" tags "+home+Lisa")
             ("hp" tags "+home+Peter")
             ("hk" tags "+home+Kim")))

The initial string in each entry defines the keys you have to press after the dispatcher command C-c a in order to access the command. Usually this will be just a single character, but if you have many similar commands, you can also define two-letter combinations where the first character is the same in several combinations and serves as a prefix key113. The second parameter is the search type, followed by the string or regular expression to be used for the matching. The example above will therefore define:

C-c a x
as a global search for agenda entries planned114 this week/day.
C-c a y
as a global search for agenda entries planned this week/day, but only those with an hour specification like [h]h:mm—think of them as appointments.
C-c a w
as a global search for TODO entries with ‘WAITING’ as the TODO keyword
C-c a W
as the same search, but only in the current buffer and displaying the results as a sparse tree
C-c a u
as a global tags search for headlines marked ‘:boss:’ but not ‘:urgent:
C-c a v
as the same search as C-c a u, but limiting the search to headlines that are also TODO items
C-c a U
as the same search as C-c a u, but only in the current buffer and displaying the result as a sparse tree
C-c a f
to create a sparse tree (again: current buffer only) with all entries containing the word ‘FIXME
C-c a h
as a prefix command for a HOME tags search where you have to press an additional key (l, p or k) to select a name (Lisa, Peter, or Kim) as additional tag to match.

Note that the *-tree agenda views need to be called from an Org buffer as they operate on the current buffer only.

Next: , Previous: Storing searches, Up: Custom agenda views

10.6.2 Block agenda

Another possibility is the construction of agenda views that comprise the results of several commands, each of which creates a block in the agenda buffer. The available commands include agenda for the daily or weekly agenda (as created with C-c a a), alltodo for the global TODO list (as constructed with C-c a t), and the matching commands discussed above: todo, tags, and tags-todo. Here are two examples:

     (setq org-agenda-custom-commands
           '(("h" "Agenda and Home-related tasks"
              ((agenda "")
               (tags-todo "home")
               (tags "garden")))
             ("o" "Agenda and Office-related tasks"
              ((agenda "")
               (tags-todo "work")
               (tags "office")))))

This will define C-c a h to create a multi-block view for stuff you need to attend to at home. The resulting agenda buffer will contain your agenda for the current week, all TODO items that carry the tag ‘home’, and also all lines tagged with ‘garden’. Finally the command C-c a o provides a similar view for office tasks.

Previous: Block agenda, Up: Custom agenda views

10.6.3 Setting options for custom commands

Org mode contains a number of variables regulating agenda construction and display. The global variables define the behavior for all agenda commands, including the custom commands. However, if you want to change some settings just for a single custom view, you can do so. Setting options requires inserting a list of variable names and values at the right spot in org-agenda-custom-commands. For example:

     (setq org-agenda-custom-commands
           '(("w" todo "WAITING"
              ((org-agenda-sorting-strategy '(priority-down))
               (org-agenda-prefix-format "  Mixed: ")))
             ("U" tags-tree "+boss-urgent"
              ((org-show-following-heading nil)
               (org-show-hierarchy-above nil)))
             ("N" search ""
              ((org-agenda-files '("~org/"))
               (org-agenda-text-search-extra-files nil)))))

Now the C-c a w command will sort the collected entries only by priority, and the prefix format is modified to just say ‘ Mixed: ’ instead of giving the category of the entry. The sparse tags tree of C-c a U will now turn out ultra-compact, because neither the headline hierarchy above the match, nor the headline following the match will be shown. The command C-c a N will do a text search limited to only a single file.

For command sets creating a block agenda, org-agenda-custom-commands has two separate spots for setting options. You can add options that should be valid for just a single command in the set, and options that should be valid for all commands in the set. The former are just added to the command entry; the latter must come after the list of command entries. Going back to the block agenda example (see Block agenda), let's change the sorting strategy for the C-c a h commands to priority-down, but let's sort the results for GARDEN tags query in the opposite order, priority-up. This would look like this:

     (setq org-agenda-custom-commands
           '(("h" "Agenda and Home-related tasks"
               (tags-todo "home")
               (tags "garden"
                     ((org-agenda-sorting-strategy '(priority-up)))))
              ((org-agenda-sorting-strategy '(priority-down))))
             ("o" "Agenda and Office-related tasks"
               (tags-todo "work")
               (tags "office")))))

As you see, the values and parentheses setting is a little complex. When in doubt, use the customize interface to set this variable—it fully supports its structure. Just one caveat: when setting options in this interface, the values are just Lisp expressions. So if the value is a string, you need to add the double-quotes around the value yourself.

To control whether an agenda command should be accessible from a specific context, you can customize org-agenda-custom-commands-contexts. Let's say for example that you have an agenda commands "o" displaying a view that you only need when reading emails. Then you would configure this option like this:

     (setq org-agenda-custom-commands-contexts
           '(("o" (in-mode . "message-mode"))))

You can also tell that the command key "o" should refer to another command key "r". In that case, add this command key like this:

     (setq org-agenda-custom-commands-contexts
           '(("o" "r" (in-mode . "message-mode"))))

See the docstring of the variable for more information.

Next: , Previous: Custom agenda views, Up: Agenda Views

10.7 Exporting Agenda Views

If you are away from your computer, it can be very useful to have a printed version of some agenda views to carry around. Org mode can export custom agenda views as plain text, HTML115, Postscript, PDF116, and iCalendar files. If you want to do this only occasionally, use the command

C-x C-w     (org-agenda-write)
Write the agenda view to a file. Depending on the extension of the selected file name, the view will be exported as HTML (extension .html or .htm), Postscript (extension .ps), iCalendar (extension .ics), or plain text (any other extension). Use the variable org-agenda-exporter-settings to set options for ps-print and for htmlize to be used during export, for example

          (setq org-agenda-exporter-settings
                '((ps-number-of-columns 2)
                  (ps-landscape-mode t)
                  (org-agenda-add-entry-text-maxlines 5)
                  (htmlize-output-type 'css)))

If you need to export certain agenda views frequently, you can associate any custom agenda command with a list of output file names 117. Here is an example that first defines custom commands for the agenda and the global TODO list, together with a number of files to which to export them. Then we define two block agenda commands and specify file names for them as well. File names can be relative to the current working directory, or absolute.

     (setq org-agenda-custom-commands
           '(("X" agenda "" nil ("agenda.html" ""))
             ("Y" alltodo "" nil ("todo.html" "todo.txt" ""))
             ("h" "Agenda and Home-related tasks"
              ((agenda "")
               (tags-todo "home")
               (tags "garden"))
             ("o" "Agenda and Office-related tasks"
               (tags-todo "work")
               (tags "office"))
              ("~/views/" "~/calendars/office.ics"))))

The extension of the file name determines the type of export. If it is .html, Org mode will use the htmlize.el package to convert the buffer to HTML and save it to this file name. If the extension is .ps, ps-print-buffer-with-faces is used to produce Postscript output. If the extension is .ics, iCalendar export is run export over all files that were used to construct the agenda, and limit the export to entries listed in the agenda. Any other extension produces a plain ASCII file.

The export files are not created when you use one of those commands interactively because this might use too much overhead. Instead, there is a special command to produce all specified files in one step:

C-c a e     (org-store-agenda-views)
Export all agenda views that have export file names associated with them.

You can use the options section of the custom agenda commands to also set options for the export commands. For example:

     (setq org-agenda-custom-commands
           '(("X" agenda ""
              ((ps-number-of-columns 2)
               (ps-landscape-mode t)
               (org-agenda-prefix-format " [ ] ")
               (org-agenda-with-colors nil)
               (org-agenda-remove-tags t))

This command sets two options for the Postscript exporter, to make it print in two columns in landscape format—the resulting page can be cut in two and then used in a paper agenda. The remaining settings modify the agenda prefix to omit category and scheduling information, and instead include a checkbox to check off items. We also remove the tags to make the lines compact, and we don't want to use colors for the black-and-white printer. Settings specified in org-agenda-exporter-settings will also apply, but the settings in org-agenda-custom-commands take precedence.

From the command line you may also use

     emacs -eval (org-batch-store-agenda-views) -kill

or, if you need to modify some parameters118

     emacs -eval '(org-batch-store-agenda-views                      \
                   org-agenda-span (quote month)                     \
                   org-agenda-start-day "2007-11-01"                 \
                   org-agenda-include-diary nil                      \
                   org-agenda-files (quote ("~/org/")))'  \

which will create the agenda views restricted to the file ~/org/, without diary entries and with a 30-day extent.

You can also extract agenda information in a way that allows further processing by other programs. See Extracting agenda information, for more information.

Previous: Exporting Agenda Views, Up: Agenda Views

10.8 Using column view in the agenda

Column view (see Column view) is normally used to view and edit properties embedded in the hierarchical structure of an Org file. It can be quite useful to use column view also from the agenda, where entries are collected by certain criteria.

C-c C-x C-c     (org-agenda-columns)
Turn on column view in the agenda.

To understand how to use this properly, it is important to realize that the entries in the agenda are no longer in their proper outline environment. This causes the following issues:

  1. Org needs to make a decision which COLUMNS format to use. Since the entries in the agenda are collected from different files, and different files may have different COLUMNS formats, this is a non-trivial problem. Org first checks if the variable org-agenda-overriding-columns-format is currently set, and if so, takes the format from there. Otherwise it takes the format associated with the first item in the agenda, or, if that item does not have a specific format (defined in a property, or in its file), it uses org-columns-default-format.
  2. If any of the columns has a summary type defined (see Column attributes), turning on column view in the agenda will visit all relevant agenda files and make sure that the computations of this property are up to date. This is also true for the special CLOCKSUM property. Org will then sum the values displayed in the agenda. In the daily/weekly agenda, the sums will cover a single day; in all other views they cover the entire block. It is vital to realize that the agenda may show the same entry twice (for example as scheduled and as a deadline), and it may show two entries from the same hierarchy (for example a parent and its child). In these cases, the summation in the agenda will lead to incorrect results because some values will count double.
  3. When the column view in the agenda shows the CLOCKSUM, that is always the entire clocked time for this item. So even in the daily/weekly agenda, the clocksum listed in column view may originate from times outside the current view. This has the advantage that you can compare these values with a column listing the planned total effort for a task—one of the major applications for column view in the agenda. If you want information about clocked time in the displayed period use clock table mode (press R in the agenda).
  4. When the column view in the agenda shows the CLOCKSUM_T, that is always today's clocked time for this item. So even in the weekly agenda, the clocksum listed in column view only originates from today. This lets you compare the time you spent on a task for today, with the time already spent (via CLOCKSUM) and with the planned total effort for it.

Next: , Previous: Agenda Views, Up: Top

11 Markup for rich export

When exporting Org mode documents, the exporter tries to reflect the structure of the document as accurately as possible in the back-end. Since export targets like HTML, LaTeX allow much richer formatting, Org mode has rules on how to prepare text for rich export. This section summarizes the markup rules used in an Org mode buffer.

Next: , Previous: Markup, Up: Markup

11.1 Structural markup elements

Next: , Previous: Structural markup elements, Up: Structural markup elements

Document title

The title of the exported document is taken from the special line

     #+TITLE: This is the title of the document

If this line does not exist, the title will be the name of the file associated to buffer, without extension, or the buffer name.

If you are exporting only a subtree, its heading will become the title of the document. If the subtree has a property EXPORT_TITLE, that will take precedence.

Next: , Previous: Document title, Up: Structural markup elements

Headings and sections

The outline structure of the document as described in Document Structure, forms the basis for defining sections of the exported document. However, since the outline structure is also used for (for example) lists of tasks, only the first three outline levels will be used as headings. Deeper levels will become itemized lists. You can change the location of this switch globally by setting the variable org-export-headline-levels, or on a per-file basis with a line

     #+OPTIONS: H:4

Next: , Previous: Headings and sections, Up: Structural markup elements

Table of contents

The table of contents is normally inserted directly before the first headline of the file. The depth of the table is by default the same as the number of headline levels, but you can choose a smaller number, or turn off the table of contents entirely, by configuring the variable org-export-with-toc, or on a per-file basis with a line like

     #+OPTIONS: toc:2          (only to two levels in TOC)
     #+OPTIONS: toc:nil        (no default TOC at all)

If you would like to move the table of contents to a different location, you should turn off the default table using org-export-with-toc or #+OPTIONS and insert #+TOC: headlines N at the desired location(s).

     #+OPTIONS: toc:nil        (no default TOC)
     #+TOC: headlines 2        (insert TOC here, with two headline levels)

Multiple #+TOC: headline lines are allowed. The same TOC keyword can also generate a list of all tables (resp. all listings) with a caption in the buffer.

     #+TOC: listings           (build a list of listings)
     #+TOC: tables             (build a list of tables)

The headline's title usually determines its corresponding entry in a table of contents. However, it is possible to specify an alternative title by setting ALT_TITLE property accordingly. It will then be used when building the table.

Next: , Previous: Table of contents, Up: Structural markup elements


Plain lists as described in Plain lists, are translated to the back-end's syntax for such lists. Most back-ends support unordered, ordered, and description lists.

Next: , Previous: Lists, Up: Structural markup elements

Paragraphs, line breaks, and quoting

Paragraphs are separated by at least one empty line. If you need to enforce a line break within a paragraph, use ‘\\’ at the end of a line.

To keep the line breaks in a region, but otherwise use normal formatting, you can use this construct, which can also be used to format poetry.

      Great clouds overhead
      Tiny black birds rise and fall
      Snow covers Emacs
          -- AlexSchroeder

When quoting a passage from another document, it is customary to format this as a paragraph that is indented on both the left and the right margin. You can include quotations in Org mode documents like this:

     Everything should be made as simple as possible,
     but not any simpler -- Albert Einstein

If you would like to center some text, do it like this:

     Everything should be made as simple as possible, \\
     but not any simpler

Next: , Previous: Paragraphs, Up: Structural markup elements

Footnote markup

Footnotes defined in the way described in Footnotes, will be exported by all back-ends. Org allows multiple references to the same note, and multiple footnotes side by side.

Next: , Previous: Footnote markup, Up: Structural markup elements

Emphasis and monospace

You can make words *bold*, /italic/, _underlined_, =verbatim= and ~code~, and, if you must, ‘+strike-through+’. Text in the code and verbatim string is not processed for Org mode specific syntax, it is exported verbatim.

To turn off fontification for marked up text, you can set org-fontify-emphasized-text to nil. To narrow down the list of available markup syntax, you can customize org-emphasis-alist. To fine tune what characters are allowed before and after the markup characters, you can tweak org-emphasis-regexp-components. Beware that changing one of the above variables will no take effect until you reload Org, for which you may need to restart Emacs.

Next: , Previous: Emphasis and monospace, Up: Structural markup elements

Horizontal rules

A line consisting of only dashes, and at least 5 of them, will be exported as a horizontal line.

Previous: Horizontal rules, Up: Structural markup elements

Comment lines

Lines starting with zero or more whitespace characters followed by one ‘#’ and a whitespace are treated as comments and will never be exported. Also entire subtrees starting with the word ‘COMMENT’ will never be exported. Finally, regions surrounded by ‘#+BEGIN_COMMENT’ ... ‘#+END_COMMENT’ will not be exported.

C-c ;
Toggle the COMMENT keyword at the beginning of an entry.

Next: , Previous: Structural markup elements, Up: Markup

11.2 Images and Tables

Both the native Org mode tables (see Tables) and tables formatted with the table.el package will be exported properly. For Org mode tables, the lines before the first horizontal separator line will become table header lines. You can use the following lines somewhere before the table to assign a caption and a label for cross references, and in the text you can refer to the object with [[tab:basic-data]] (see Internal links):

     #+CAPTION: This is the caption for the next table (or link)
     #+NAME:   tab:basic-data
        | ... | ...|

Optionally, the caption can take the form:

     #+CAPTION[Caption for list of tables]: Caption for table.

Some back-ends allow you to directly include images into the exported document. Org does this, if a link to an image files does not have a description part, for example [[./img/a.jpg]]. If you wish to define a caption for the image and maybe a label for internal cross references, make sure that the link is on a line by itself and precede it with #+CAPTION and #+NAME as follows:

     #+CAPTION: This is the caption for the next figure link (or table)
     #+NAME:   fig:SED-HR4049

Such images can be displayed within the buffer. See the discussion of image links.

Even though images and tables are prominent examples of captioned structures, the same caption mechanism can apply to many others (e.g., LaTeX equations, source code blocks). Depending on the export back-end, those may or may not be handled.

Next: , Previous: Images and tables, Up: Markup

11.3 Literal examples

You can include literal examples that should not be subjected to markup. Such examples will be typeset in monospace, so this is well suited for source code and similar examples.

     Some example from a text file.

Note that such blocks may be indented in order to align nicely with indented text and in particular with plain list structure (see Plain lists). For simplicity when using small examples, you can also start the example lines with a colon followed by a space. There may also be additional whitespace before the colon:

     Here is an example
        : Some example from a text file.

If the example is source code from a programming language, or any other text that can be marked up by font-lock in Emacs, you can ask for the example to look like the fontified Emacs buffer119. This is done with the ‘src’ block, where you also need to specify the name of the major mode that should be used to fontify the example120, see Easy Templates for shortcuts to easily insert code blocks.

     #+BEGIN_SRC emacs-lisp
       (defun org-xor (a b)
          "Exclusive or."
          (if a (not b) b))

Both in example and in src snippets, you can add a -n switch to the end of the BEGIN line, to get the lines of the example numbered. If you use a +n switch, the numbering from the previous numbered snippet will be continued in the current one. In literal examples, Org will interpret strings like ‘(ref:name)’ as labels, and use them as targets for special hyperlinks like [[(name)]] (i.e., the reference name enclosed in single parenthesis). In HTML, hovering the mouse over such a link will remote-highlight the corresponding code line, which is kind of cool.

You can also add a -r switch which removes the labels from the source code121. With the -n switch, links to these references will be labeled by the line numbers from the code listing, otherwise links will use the labels with no parentheses. Here is an example:

     #+BEGIN_SRC emacs-lisp -n -r
     (save-excursion                  (ref:sc)
        (goto-char (point-min)))      (ref:jump)
     In line [[(sc)]] we remember the current position.  [[(jump)][Line (jump)]]
     jumps to point-min.

If the syntax for the label format conflicts with the language syntax, use a -l switch to change the format, for example ‘#+BEGIN_SRC pascal -n -r -l "((%s))"’. See also the variable org-coderef-label-format.

HTML export also allows examples to be published as text areas (see Text areas in HTML export).

Because the #+BEGIN_... and #+END_... patterns need to be added so often, shortcuts are provided using the Easy Templates facility (see Easy Templates).

C-c '
Edit the source code example at point in its native mode. This works by switching to a temporary buffer with the source code. You need to exit by pressing C-c ' again122. The edited version will then replace the old version in the Org buffer. Fixed-width regions (where each line starts with a colon followed by a space) will be edited using artist-mode123 to allow creating ASCII drawings easily. Using this command in an empty line will create a new fixed-width region.
C-c l
Calling org-store-link while editing a source code example in a temporary buffer created with C-c ' will prompt for a label. Make sure that it is unique in the current buffer, and insert it with the proper formatting like ‘(ref:label)’ at the end of the current line. Then the label is stored as a link ‘(label)’, for retrieval with C-c C-l.

Next: , Previous: Literal examples, Up: Markup

11.4 Include files

During export, you can include the content of another file. For example, to include your .emacs file, you could use:

     #+INCLUDE: "~/.emacs" src emacs-lisp

The optional second and third parameter are the markup (i.e., ‘example’ or ‘src’), and, if the markup is ‘src’, the language for formatting the contents. The markup is optional; if it is not given, the text will be assumed to be in Org mode format and will be processed normally.

Contents of the included file will belong to the same structure (headline, item) containing the INCLUDE keyword. In particular, headlines within the file will become children of the current section. That behavior can be changed by providing an additional keyword parameter, :minlevel. In that case, all headlines in the included file will be shifted so the one with the lowest level reaches that specified level. For example, to make a file become a sibling of the current top-level headline, use

     #+INCLUDE: "~/my-book/" :minlevel 1

You can also include a portion of a file by specifying a lines range using the :lines parameter. The line at the upper end of the range will not be included. The start and/or the end of the range may be omitted to use the obvious defaults.

     #+INCLUDE: "~/.emacs" :lines "5-10"   Include lines 5 to 10, 10 excluded
     #+INCLUDE: "~/.emacs" :lines "-10"    Include lines 1 to 10, 10 excluded
     #+INCLUDE: "~/.emacs" :lines "10-"    Include lines from 10 to EOF
C-c '
Visit the include file at point.

Next: , Previous: Include files, Up: Markup

11.5 Index entries

You can specify entries that will be used for generating an index during publishing. This is done by lines starting with #+INDEX. An entry the contains an exclamation mark will create a sub item. See Generating an index for more information.

     * Curriculum Vitae
     #+INDEX: CV
     #+INDEX: Application!CV

Next: , Previous: Index entries, Up: Markup

11.6 Macro replacement

You can define text snippets with

     #+MACRO: name   replacement text $1, $2 are arguments

which can be referenced in paragraphs, verse blocks, table cells and some keywords with {{{name(arg1,arg2)}}}124. In addition to defined macros, {{{title}}}, {{{author}}}, etc., will reference information set by the #+TITLE:, #+AUTHOR:, and similar lines. Also, {{{time(FORMAT)}}} and {{{modification-time(FORMAT)}}} refer to current date time and to the modification time of the file being exported, respectively. FORMAT should be a format string understood by format-time-string.

Macro expansion takes place during export.

Next: , Previous: Macro replacement, Up: Markup

11.7 Embedded LaTeX

Plain ASCII is normally sufficient for almost all note taking. Exceptions include scientific notes, which often require mathematical symbols and the occasional formula. LaTeX125 is widely used to typeset scientific documents. Org mode supports embedding LaTeX code into its files, because many academics are used to writing and reading LaTeX source code, and because it can be readily processed to produce pretty output for a number of export back-ends.

Next: , Previous: Embedded LaTeX, Up: Embedded LaTeX

11.7.1 Special symbols

You can use LaTeX-like syntax to insert special symbols like ‘\alpha’ to indicate the Greek letter, or ‘\to’ to indicate an arrow. Completion for these symbols is available, just type ‘\’ and maybe a few letters, and press M-<TAB> to see possible completions. Unlike LaTeX code, Org mode allows these symbols to be present without surrounding math delimiters, for example:

     Angles are written as Greek letters \alpha, \beta and \gamma.

During export, these symbols will be transformed into the native format of the exporter back-end. Strings like \alpha will be exported as &alpha; in the HTML output, and as $\alpha$ in the LaTeX output. Similarly, \nbsp will become &nbsp; in HTML and ~ in LaTeX. If you need such a symbol inside a word, terminate it like this: ‘\Aacute{}stor’.

A large number of entities is provided, with names taken from both HTML and LaTeX; see the variable org-entities for the complete list. ‘\-’ is treated as a shy hyphen, and ‘--’, ‘---’, and ‘...’ are all converted into special commands creating hyphens of different lengths or a compact set of dots.

If you would like to see entities displayed as UTF-8 characters, use the following command126:

C-c C-x \
Toggle display of entities as UTF-8 characters. This does not change the buffer content which remains plain ASCII, but it overlays the UTF-8 character for display purposes only.

Next: , Previous: Special symbols, Up: Embedded LaTeX

11.7.2 Subscripts and superscripts

Just like in LaTeX, ‘^’ and ‘_’ are used to indicate super- and subscripts. Again, these can be used without embedding them in math-mode delimiters. To increase the readability of ASCII text, it is not necessary (but OK) to surround multi-character sub- and superscripts with curly braces. For example

     The mass of the sun is M_sun = 1.989 x 10^30 kg.  The radius of
     the sun is R_{sun} = 6.96 x 10^8 m.

If you write a text where the underscore is often used in a different context, Org's convention to always interpret these as subscripts can get in your way. Configure the variable org-use-sub-superscripts to change this convention. For example, when setting this variable to {}, ‘a_b’ will not be interpreted as a subscript, but ‘a_{b}’ will.

C-c C-x \
In addition to showing entities as UTF-8 characters, this command will also format sub- and superscripts in a WYSIWYM way.

Next: , Previous: Subscripts and superscripts, Up: Embedded LaTeX

11.7.3 LaTeX fragments

Going beyond symbols and sub- and superscripts, a full formula language is needed. Org mode can contain LaTeX math fragments, and it supports ways to process these for several export back-ends. When exporting to LaTeX, the code is obviously left as it is. When exporting to HTML, Org invokes the MathJax library (see Math formatting in HTML export) to process and display the math127. Finally, it can also process the mathematical expressions into images128 that can be displayed in a browser.

LaTeX fragments don't need any special marking at all. The following snippets will be identified as LaTeX source code:

For example:

     If $a^2=b$ and \( b=2 \), then the solution must be
     either $$ a=+\sqrt{2} $$ or \[ a=-\sqrt{2} \].

LaTeX processing can be configured with the variable org-export-with-latex. The default setting is t which means MathJax for HTML, and no processing for ASCII and LaTeX back-ends. You can also set this variable on a per-file basis using one of these lines:

     #+OPTIONS: tex:t          Do the right thing automatically (MathJax)
     #+OPTIONS: tex:nil        Do not process LaTeX fragments at all
     #+OPTIONS: tex:verbatim   Verbatim export, for jsMath or so

Next: , Previous: LaTeX fragments, Up: Embedded LaTeX

11.7.4 Previewing LaTeX fragments

If you have dvipng or imagemagick installed130, LaTeX fragments can be processed to produce preview images of the typeset expressions:

C-c C-x C-l
Produce a preview image of the LaTeX fragment at point and overlay it over the source code. If there is no fragment at point, process all fragments in the current entry (between two headlines). When called with a prefix argument, process the entire subtree. When called with two prefix arguments, or when the cursor is before the first headline, process the entire buffer.
C-c C-c
Remove the overlay preview images.

You can customize the variable org-format-latex-options to influence some aspects of the preview. In particular, the :scale (and for HTML export, :html-scale) property can be used to adjust the size of the preview images.

You can turn on the previewing of all LaTeX fragments in a file with

     #+STARTUP: latexpreview

To disable it, simply use

     #+STARTUP: nolatexpreview

Previous: Previewing LaTeX fragments, Up: Embedded LaTeX

11.7.5 Using CDLaTeX to enter math

CDLaTeX mode is a minor mode that is normally used in combination with a major LaTeX mode like AUCTeX in order to speed-up insertion of environments and math templates. Inside Org mode, you can make use of some of the features of CDLaTeX mode. You need to install cdlatex.el and texmathp.el (the latter comes also with AUCTeX) from Don't use CDLaTeX mode itself under Org mode, but use the light version org-cdlatex-mode that comes as part of Org mode. Turn it on for the current buffer with M-x org-cdlatex-mode RET, or for all Org files with

     (add-hook 'org-mode-hook 'turn-on-org-cdlatex)

When this mode is enabled, the following features are present (for more details see the documentation of CDLaTeX mode):

Previous: Embedded LaTeX, Up: Markup

11.8 Special blocks

Org syntax includes pre-defined blocks (see Paragraphs and Literal examples). It is also possible to create blocks containing raw code targeted at a specific back-ends (e.g., ‘#+BEGIN_LATEX’).

Any other block is a special block.

For example, ‘#+BEGIN_ABSTRACT’ and ‘#+BEGIN_VIDEO’ are special blocks. The first one is useful when exporting to LaTeX, the second one when exporting to HTML5.

Each export back-end decides if they should be exported, and how. When the block is ignored, its contents are still exported, as if the opening and closing block lines were not there. For example, when exporting a ‘#+BEGIN_TEST’ block, HTML back-end wraps its contents within a ‘<div name="test">’ tag.

Refer to back-end specific documentation for more information.

Next: , Previous: Markup, Up: Top

12 Exporting

The Org mode export facilities can be used to export Org documents or parts of Org documents to a variety of other formats. In addition, these facilities can be used with orgtbl-mode and/or orgstruct-mode in foreign buffers so you can author tables and lists in Org syntax and convert them in place to the target language.

ASCII export produces a readable and simple version of an Org file for printing and sharing notes. HTML export allows you to easily publish notes on the web, or to build full-fledged websites. LaTeX export lets you use Org mode and its structured editing functions to create arbitrarily complex LaTeX files for any kind of document. OpenDocument Text (ODT) export allows seamless collaboration across organizational boundaries. Markdown export lets you seamlessly collaborate with other developers. Finally, iCal export can extract entries with deadlines or appointments to produce a file in the iCalendar format.

Next: , Previous: Exporting, Up: Exporting

12.1 The Export Dispatcher

The main entry point for export related tasks is the dispatcher, a hierarchical menu from which it is possible to select an export format and toggle export options132 from which it is possible to select an export format and to toggle export options.

C-c C-e     (org-export-dispatch)

Dispatch for export and publishing commands. When called with a C-u prefix argument, repeat the last export command on the current buffer while preserving toggled options. If the current buffer hasn't changed and subtree export was activated, the command will affect that same subtree.

Normally the entire buffer is exported, but if there is an active region only that part of the buffer will be exported.

Several export options (see Export settings) can be toggled from the export dispatcher with the following key combinations:

Toggle asynchronous export. Asynchronous export uses an external Emacs process that is configured with a specified initialization file.

While exporting asynchronously, the output is not displayed, but stored in a place called “the export stack”. This stack can be displayed by calling the dispatcher with a double C-u prefix argument, or with & key from the dispatcher menu.

To make this behavior the default, customize the variable org-export-in-background.

Toggle body-only export. Its effect depends on the back-end used. Typically, if the back-end has a header section (like <head>...</head> in the HTML back-end), a body-only export will not include this header.
Toggle subtree export. The top heading becomes the document title.

You can change the default state of this option by setting org-export-initial-scope.

Toggle visible-only export. Only export the text that is currently visible, i.e., not hidden by outline visibility in the buffer.

With the exception of asynchronous export, a successful export process writes its output to the kill-ring. You can configure this behavior by altering the option org-export-copy-to-kill-ring.

Next: , Previous: The Export Dispatcher, Up: Exporting

12.2 Export back-ends

An export back-end is a library that translates Org syntax into a foreign format. An export format is not available until the proper back-end has been loaded.

By default, the following four back-ends are loaded: ascii, html, icalendar and latex. It is possible to add more (or remove some) by customizing org-export-backends.

Built-in back-ends include:

Other back-ends might be found in the contrib/ directory (see Installation).

Next: , Previous: Export back-ends, Up: Exporting

12.3 Export settings

Export options can be set: globally with variables; for an individual file by making variables buffer-local with in-buffer settings (see In-buffer settings), by setting individual keywords, or by specifying them in a compact form with the #+OPTIONS keyword; or for a tree by setting properties (see Properties and Columns). Options set at a specific level override options set at a more general level.

In-buffer settings may appear anywhere in the file, either directly or indirectly through a file included using ‘#+SETUPFILE: filename’ syntax. Option keyword sets tailored to a particular back-end can be inserted from the export dispatcher (see The Export Dispatcher) using the Insert template command by pressing <#>. To insert keywords individually, a good way to make sure the keyword is correct is to type #+ and then to use M-<TAB> for completion.

The export keywords available for every back-end, and their equivalent global variables, include:

The document author (user-full-name).
Entity responsible for output generation (org-export-creator-string).
A date or a time-stamp133.
The document description. Back-ends handle it as they see fit (e.g., for the XHTML meta tag), if at all. You can use several such keywords for long descriptions.
The email address (user-mail-address).
The keywords defining the contents of the document. Back-ends handle it as they see fit (e.g., for the XHTML meta tag), if at all. You can use several such keywords if the list is long.
The language used for translating some strings (org-export-default-language). E.g., ‘#+LANGUAGE: fr’ will tell Org to translate File (english) into Fichier (french) in the clocktable.
The tags that select a tree for export (org-export-select-tags). The default value is :export:. Within a subtree tagged with :export:, you can still exclude entries with :noexport: (see below). When headlines are selectively exported with :export: anywhere in a file, text before the first headline is ignored.
The tags that exclude a tree from export (org-export-exclude-tags). The default value is :noexport:. Entries with the :noexport: tag will be unconditionally excluded from the export, even if they have an :export: tag.
The title to be shown (otherwise derived from buffer's name). You can use several such keywords for long titles.

The #+OPTIONS keyword is a compact134 form that recognizes the following arguments:

Toggle smart quotes (org-export-with-smart-quotes).
Toggle emphasized text (org-export-with-emphasize).
Toggle conversion of special strings (org-export-with-special-strings).
Toggle fixed-width sections (org-export-with-fixed-width).
Toggle inclusion of any time/date active/inactive stamps (org-export-with-timestamps).
Toggle line-break-preservation (org-export-preserve-breaks).
Toggle TeX-like syntax for sub- and superscripts. If you write "^:{}", ‘a_{b}’ will be interpreted, but the simple ‘a_b’ will be left as it is (org-export-with-sub-superscripts).
Configure export of archived trees. Can be set to headline to only process the headline, skipping its contents (org-export-with-archived-trees).
Toggle inclusion of author name into exported file (org-export-with-author).
Toggle inclusion of CLOCK keywords (org-export-with-clocks).
Configure inclusion of creator info into exported file. It may be set to comment (org-export-with-creator).
Toggle inclusion of drawers, or list drawers to include (org-export-with-drawers).
Toggle inclusion of entities (org-export-with-entities).
Toggle inclusion of the author's e-mail into exported file (org-export-with-email).
Toggle the inclusion of footnotes (org-export-with-footnotes).
Set the number of headline levels for export (org-export-headline-levels). Below that level, headlines are treated differently. In most back-ends, they become list items.
Toggle inclusion of inlinetasks (org-export-with-inlinetasks).
Toggle section-numbers (org-export-with-section-numbers). It can also be set to a number ‘n’, so only headlines at that level or above will be numbered.
Toggle export of planning information (org-export-with-planning). “Planning information” is the line containing the SCHEDULED:, the DEADLINE: or the CLOSED: cookies or a combination of them.
Toggle inclusion of priority cookies (org-export-with-priority).
Toggle inclusion of statistics cookies (org-export-with-statistics-cookies).
Toggle inclusion of tags, may also be not-in-toc (org-export-with-tags).
Toggle inclusion of tasks (TODO items), can be nil to remove all tasks, todo to remove DONE tasks, or a list of keywords to keep (org-export-with-tasks).
Configure export of LaTeX fragments and environments. It may be set to verbatim (org-export-with-latex).
Toggle inclusion of the creation time into exported file (org-export-time-stamp-file).
Toggle inclusion of the table of contents, or set the level limit (org-export-with-toc).
Toggle inclusion of TODO keywords into exported text (org-export-with-todo-keywords).
Toggle inclusion of tables (org-export-with-tables).

When exporting only a subtree, each of the previous keywords135 can be overridden locally by special node properties. These begin with ‘EXPORT_’, followed by the name of the keyword they supplant. For example, ‘DATE’ and ‘OPTIONS’ keywords become, respectively, ‘EXPORT_DATE’ and ‘EXPORT_OPTIONS’ properties.

If org-export-allow-bind-keywords is non-nil, Emacs variables can become buffer-local during export by using the BIND keyword. Its syntax is ‘#+BIND: variable value’. This is particularly useful for in-buffer settings that cannot be changed using specific keywords.

The name of the output file to be generated is taken from the file associated to the buffer, when possible, or asked to you otherwise. For subtree export, you can also set ‘EXPORT_FILE_NAME’ property. In all cases, only the base name of the file is retained, and a back-end specific extension is added.

Next: , Previous: Export settings, Up: Exporting

12.4 ASCII/Latin-1/UTF-8 export

ASCII export produces a simple and very readable version of an Org mode file, containing only plain ASCII. Latin-1 and UTF-8 export augment the file with special characters and symbols available in these encodings.

Links are exported in a footnote-like style, with the descriptive part in the text and the link in a note before the next heading. See the variable org-ascii-links-to-notes for details and other options.

ASCII export commands

C-c C-e t a/l/u     (org-ascii-export-to-ascii)
Export as an ASCII file. For an Org file,, the ASCII file will be myfile.txt. The file will be overwritten without warning. When the original file is myfile.txt, the resulting file becomes myfile.txt.txt in order to prevent data loss.
C-c C-e t A/L/U     (org-ascii-export-as-ascii)
Export to a temporary buffer. Do not create a file.

Header and sectioning structure

In the exported version, the first three outline levels become headlines, defining a general document structure. Additional levels are exported as lists. The transition can also occur at a different level (see Export settings).

Quoting ASCII text

You can insert text that will only appear when using ASCII back-end with the following constructs:

     Text @@ascii:and additional text@@ within a paragraph.
     #+ASCII: Some text
     All lines in this block will appear only when using this back-end.

ASCII specific attributes

ASCII back-end only understands one attribute, :width, which specifies the length, in characters, of a given horizontal rule. It must be specified using an ATTR_ASCII line, directly preceding the rule.

     #+ATTR_ASCII: :width 10

Next: , Previous: ASCII/Latin-1/UTF-8 export, Up: Exporting

12.5 Beamer export

The LaTeX class Beamer allows production of high quality presentations using LaTeX and pdf processing. Org mode has special support for turning an Org mode file or tree into a Beamer presentation.

Beamer export commands

C-c C-e l b     (org-beamer-export-to-latex)
Export as a LaTeX file. For an Org file, the LaTeX file will be myfile.tex. The file will be overwritten without warning.
C-c C-e l B     (org-beamer-export-as-latex)
Export to a temporary buffer. Do not create a file.
C-c C-e l P     (org-beamer-export-to-pdf)
Export as LaTeX and then process to PDF.
C-c C-e l O
Export as LaTeX and then process to PDF, then open the resulting PDF file.

Sectioning, Frames and Blocks

Any tree with not-too-deep level nesting should in principle be exportable as a Beamer presentation. Headlines fall into three categories: sectioning elements, frames and blocks.

Headlines also support BEAMER_ACT and BEAMER_OPT properties. The former is translated as an overlay/action specification, or a default overlay specification when enclosed within square brackets. The latter specifies options137 for the current frame or block. The export back-end will automatically wrap properties within angular or square brackets when appropriate.

Moreover, headlines handle the BEAMER_COL property. Its value should be a decimal number representing the width of the column as a fraction of the total text width. If the headline has no specific environment, its title will be ignored and its contents will fill the column created. Otherwise, the block will fill the whole column and the title will be preserved. Two contiguous headlines with a non-nil BEAMER_COL value share the same columns LaTeX environment. It will end before the next headline without such a property. This environment is generated automatically. Although, it can also be explicitly created, with a special columns value for BEAMER_ENV property (if it needs to be set up with some specific options, for example).

Beamer specific syntax

Beamer back-end is an extension of LaTeX back-end. As such, all LaTeX specific syntax (e.g., ‘#+LATEX:’ or ‘#+ATTR_LATEX:’) is recognized. See LaTeX and PDF export for more information.

Beamer export introduces a number of keywords to insert code in the document's header. Four control appearance of the presentation: #+BEAMER_THEME, #+BEAMER_COLOR_THEME, #+BEAMER_FONT_THEME, #+BEAMER_INNER_THEME and #+BEAMER_OUTER_THEME. All of them accept optional arguments within square brackets. The last one, #+BEAMER_HEADER, is more generic and allows you to append any line of code in the header.

     #+BEAMER_THEME: Rochester [height=20pt]
     #+BEAMER_COLOR_THEME: spruce

Table of contents generated from toc:t OPTION keyword are wrapped within a frame environment. Those generated from a TOC keyword (see Table of contents) are not. In that case, it is also possible to specify options, enclosed within square brackets.

     #+TOC: headlines [currentsection]

Beamer specific code can be inserted with the following constructs:

     #+BEAMER: \pause
     All lines in this block will appear only when using this back-end.
     Text @@beamer:some code@@ within a paragraph.

In particular, this last example can be used to add overlay specifications to objects whose type is among bold, item, link, radio-target and target, when the value is enclosed within angular brackets and put at the beginning the object.

     A *@@beamer:<2->@@useful* feature

Eventually, every plain list has support for :environment, :overlay and :options attributes through ATTR_BEAMER affiliated keyword. The first one allows the use of a different environment, the second sets overlay specifications and the last one inserts optional arguments in current list environment.

     #+ATTR_BEAMER: :overlay +-
     - item 1
     - item 2

Editing support

You can turn on a special minor mode org-beamer-mode for faster editing with:

     #+STARTUP: beamer
C-c C-b     (org-beamer-select-environment)
In org-beamer-mode, this key offers fast selection of a Beamer environment or the BEAMER_COL property.

Also, a template for useful in-buffer settings or properties can be inserted into the buffer with M-x org-beamer-insert-options-template. Among other things, this will install a column view format which is very handy for editing special properties used by Beamer.

An example

Here is a simple example Org document that is intended for Beamer export.

     #+TITLE: Example Presentation
     #+AUTHOR: Carsten Dominik
     #+OPTIONS: H:2
     #+LATEX_CLASS: beamer
     #+LATEX_CLASS_OPTIONS: [presentation]
     #+BEAMER_THEME: Madrid
     * This is the first structural section
     ** Frame 1
     *** Thanks to Eric Fraga                                           :B_block:BMCOL:
         :BEAMER_COL: 0.48
         :BEAMER_ENV: block
         for the first viable Beamer setup in Org
     *** Thanks to everyone else                                        :B_block:BMCOL:
         :BEAMER_COL: 0.48
         :BEAMER_ACT: <2->
         :BEAMER_ENV: block
         for contributing to the discussion
     **** This will be formatted as a beamer note                              :B_note:
          :BEAMER_env: note
     ** Frame 2 (where we will not use columns)
     *** Request
         Please test this stuff!

Next: , Previous: Beamer export, Up: Exporting

12.6 HTML export

Org mode contains an HTML (XHTML 1.0 strict) exporter with extensive HTML formatting, in ways similar to John Gruber's markdown language, but with additional support for tables.

Next: , Previous: HTML export, Up: HTML export

12.6.1 HTML export commands

C-c C-e h h     (org-html-export-to-html)
Export as an HTML file. For an Org file, the HTML file will be myfile.html. The file will be overwritten without warning. C-c C-e h o Export as an HTML file and immediately open it with a browser.
C-c C-e h H     (org-html-export-as-html)
Export to a temporary buffer. Do not create a file.

Next: , Previous: HTML Export commands, Up: HTML export

12.6.2 HTML doctypes

Org can export to various (X)HTML flavors.

Setting the variable org-html-doctype allows you to export to different (X)HTML variants. The exported HTML will be adjusted according to the syntax requirements of that variant. You can either set this variable to a doctype string directly, in which case the exporter will try to adjust the syntax automatically, or you can use a ready-made doctype. The ready-made options are:

See the variable org-html-doctype-alist for details. The default is “xhtml-strict”.

Fancy HTML5 export

HTML5 introduces several new element types. By default, Org will not make use of these element types, but you can set org-html-html5-fancy to t (or set html5-fancy item in an OPTIONS line), to enable a few new block-level elements. These are created using arbitrary #+BEGIN and #+END blocks. For instance:

     Lorem ipsum

Will export to:

       <p>Lorem ipsum</p>

While this:

     #+ATTR_HTML: :controls controls :width 350
     #+HTML: <source src="movie.mp4" type="video/mp4">
     #+HTML: <source src="movie.ogg" type="video/ogg">
     Your browser does not support the video tag.


     <video controls="controls" width="350">
       <source src="movie.mp4" type="video/mp4">
       <source src="movie.ogg" type="video/ogg">
       <p>Your browser does not support the video tag.</p>

Special blocks that do not correspond to HTML5 elements (see org-html-html5-elements) will revert to the usual behavior, i.e., #+BEGIN_LEDERHOSEN will still export to ‘<div class="lederhosen">’.

Headlines cannot appear within special blocks. To wrap a headline and its contents in e.g., ‘<section>’ or ‘<article>’ tags, set the HTML_CONTAINER property on the headline itself.

Next: , Previous: HTML doctypes, Up: HTML export

12.6.3 HTML preamble and postamble

The HTML exporter lets you define a preamble and a postamble.

The default value for org-html-preamble is t, which means that the preamble is inserted depending on the relevant format string in org-html-preamble-format.

Setting org-html-preamble to a string will override the default format string. If you set it to a function, it will insert the output of the function, which must be a string. Setting to nil will not insert any preamble.

The default value for org-html-postamble is 'auto, which means that the HTML exporter will look for information about the author, the email, the creator and the date, and build the postamble from these values. Setting org-html-postamble to t will insert the postamble from the relevant format string found in org-html-postamble-format. Setting it to nil will not insert any postamble.

Next: , Previous: HTML preamble and postamble, Up: HTML export

12.6.4 Quoting HTML tags

Plain ‘<’ and ‘>’ are always transformed to ‘&lt;’ and ‘&gt;’ in HTML export. If you want to include raw HTML code, which should only appear in HTML export, mark it with ‘@@html:’ as in ‘@@html:<b>@@bold text@@html:</b>@@’. For more extensive HTML that should be copied verbatim to the exported file use either

     #+HTML: Literal HTML code for export


     All lines between these markers are exported literally

Next: , Previous: Quoting HTML tags, Up: HTML export

12.6.5 Links in HTML export

Internal links (see Internal links) will continue to work in HTML. This includes automatic links created by radio targets (see Radio targets). Links to external files will still work if the target file is on the same relative path as the published Org file. Links to other .org files will be translated into HTML links under the assumption that an HTML version also exists of the linked file, at the same relative path. ‘id:’ links can then be used to jump to specific entries across files. For information related to linking files while publishing them to a publishing directory see Publishing links.

If you want to specify attributes for links, you can do so using a special #+ATTR_HTML line to define attributes that will be added to the <a> or <img> tags. Here is an example that sets title and style attributes for a link:

     #+ATTR_HTML: :title The Org mode homepage :style color:red;

Next: , Previous: Links in HTML export, Up: HTML export

12.6.6 Tables

Org mode tables are exported to HTML using the table attributes defined in org-html-table-default-attributes. The default setting makes tables without cell borders and frame. If you would like to change this for individual tables, place something like the following before the table:

     #+CAPTION: This is a table with lines around and between cells
     #+ATTR_HTML: :border 2 :rules all :frame border

You can also modify the default tags used for each row by setting org-html-table-row-tags. See the docstring for an example on how to use this option.

Next: , Previous: Tables in HTML export, Up: HTML export

12.6.7 Images in HTML export

HTML export can inline images given as links in the Org file, and it can make an image the clickable part of a link. By default138, images are inlined if a link does not have a description. So ‘[[file:myimg.jpg]]’ will be inlined, while ‘[[file:myimg.jpg][the image]]’ will just produce a link ‘the image’ that points to the image. If the description part itself is a file: link or a http: URL pointing to an image, this image will be inlined and activated so that clicking on the image will activate the link. For example, to include a thumbnail that will link to a high resolution version of the image, you could use:


If you need to add attributes to an inlined image, use a #+ATTR_HTML. In the example below we specify the alt and title attributes to support text viewers and accessibility, and align it to the right.

     #+CAPTION: A black cat stalking a spider
     #+ATTR_HTML: :alt cat/spider image :title Action! :align right

You could use http addresses just as well.

Next: , Previous: Images in HTML export, Up: HTML export

12.6.8 Math formatting in HTML export

LaTeX math snippets (see LaTeX fragments) can be displayed in two different ways on HTML pages. The default is to use the MathJax system which should work out of the box with Org mode installation because serves MathJax for Org mode users for small applications and for testing purposes. If you plan to use this regularly or on pages with significant page views, you should install139 MathJax on your own server in order to limit the load of our server. To configure MathJax, use the variable org-html-mathjax-options or insert something like the following into the buffer:

     #+HTML_MATHJAX: align:"left" mathml:t path:"/MathJax/MathJax.js"

See the docstring of the variable org-html-mathjax-options for the meaning of the parameters in this line.

If you prefer, you can also request that LaTeX fragments are processed into small images that will be inserted into the browser page. Before the availability of MathJax, this was the default method for Org files. This method requires that the dvipng program or imagemagick suite is available on your system. You can still get this processing with

     #+OPTIONS: tex:dvipng


     #+OPTIONS: tex:imagemagick

Next: , Previous: Math formatting in HTML export, Up: HTML export

12.6.9 Text areas in HTML export

An alternative way to publish literal code examples in HTML is to use text areas, where the example can even be edited before pasting it into an application. It is triggered by :textarea attribute at an example or src block.

You may also use :height and :width attributes to specify the height and width of the text area, which default to the number of lines in the example, and 80, respectively. For example

     #+ATTR_HTML: :textarea t :width 40
       (defun org-xor (a b)
          "Exclusive or."
          (if a (not b) b))

Next: , Previous: Text areas in HTML export, Up: HTML export

12.6.10 CSS support

You can modify the CSS style definitions for the exported file. The HTML exporter assigns the following special CSS classes140 to appropriate parts of the document—your style specifications may change these, in addition to any of the standard classes like for headlines, tables, etc.            author information, including email              publishing date
     p.creator           creator info, about org mode version
     .title              document title
     .todo               TODO keywords, all not-done states
     .done               the DONE keywords, all states that count as done
     .WAITING            each TODO keyword also uses a class named after itself
     .timestamp          timestamp
     .timestamp-kwd      keyword associated with a timestamp, like SCHEDULED
     .timestamp-wrapper  span around keyword plus timestamp
     .tag                tag in a headline
     ._HOME              each tag uses itself as a class, "@" replaced by "_"
     .target             target for links
     .linenr             the line number in a code example
     .code-highlighted   for highlighting referenced code lines
     div.outline-N       div for outline level N (headline plus text))
     div.outline-text-N  extra div for text at outline level N
     .section-number-N   section number in headlines, different for each level
     .figure-number      label like "Figure 1:"
     .table-number       label like "Table 1:"
     .listing-number     label like "Listing 1:"
     div.figure          how to format an inlined image
     pre.src             formatted source code
     pre.example         normal example
     p.verse             verse paragraph
     div.footnotes       footnote section headline
     p.footnote          footnote definition paragraph, containing a footnote
     .footref            a footnote reference number (always a <sup>)
     .footnum            footnote number in footnote definition (always <sup>)

Each exported file contains a compact default style that defines these classes in a basic way141. You may overwrite these settings, or add to them by using the variables org-html-head and org-html-head-extra. You can override the global values of these variables for each file by using these keywords:

     #+HTML_HEAD: <link rel="stylesheet" type="text/css" href="style1.css" />
     #+HTML_HEAD_EXTRA: <link rel="alternate stylesheet" type="text/css" href="style2.css" />

For longer style definitions, you can use several such lines. You could also directly write a <style> </style> section in this way, without referring to an external file.

In order to add styles to a subtree, use the :HTML_CONTAINER_CLASS: property to assign a class to the tree. In order to specify CSS styles for a particular headline, you can use the id specified in a :CUSTOM_ID: property.

Previous: CSS support, Up: HTML export

12.6.11 JavaScript supported display of web pages

Sebastian Rose has written a JavaScript program especially designed to enhance the web viewing experience of HTML files created with Org. This program allows you to view large files in two different ways. The first one is an Info-like mode where each section is displayed separately and navigation can be done with the n and p keys (and some other keys as well, press ? for an overview of the available keys). The second view type is a folding view much like Org provides inside Emacs. The script is available at and you can find the documentation for it at We host the script at our site, but if you use it a lot, you might not want to be dependent on and prefer to install a local copy on your own web server.

All it then takes to use this program is adding a single line to the Org file:

     #+INFOJS_OPT: view:info toc:nil

If this line is found, the HTML header will automatically contain the code needed to invoke the script. Using the line above, you can set the following viewing options:

     path:    The path to the script.  The default is to grab the script from
    , but you might want to have
              a local copy and use a path like ‘../scripts/org-info.js’.
     view:    Initial view when the website is first shown.  Possible values are:
              info      Info-like interface with one section per page.
              overview  Folding interface, initially showing only top-level.
              content   Folding interface, starting with all headlines visible.
              showall   Folding interface, all headlines and text visible.
     sdepth:  Maximum headline level that will still become an independent
              section for info and folding modes.  The default is taken from
              org-export-headline-levels (= the H switch in #+OPTIONS).
              If this is smaller than in org-export-headline-levels, each
              info/folding section can still contain child headlines.
     toc:     Should the table of contents initially be visible?
              Even when nil, you can always get to the "toc" with i.
     tdepth:  The depth of the table of contents.  The defaults are taken from
              the variables org-export-headline-levels and org-export-with-toc.
     ftoc:    Does the CSS of the page specify a fixed position for the "toc"?
              If yes, the toc will never be displayed as a section.
     ltoc:    Should there be short contents (children) in each section?
              Make this above if the section should be above initial text.
     mouse:   Headings are highlighted when the mouse is over them.  Should be
              underline’ (default) or a background color like ‘#cccccc’.
     buttons: Should view-toggle buttons be everywhere?  When nil (the
              default), only one such button will be present.

You can choose default values for these options by customizing the variable org-html-infojs-options. If you always want to apply the script to your pages, configure the variable org-html-use-infojs.

Next: , Previous: HTML export, Up: Exporting

12.7 LaTeX and PDF export

LaTeX export can produce an arbitrarily complex LaTeX document of any standard or custom document class. With further processing142, which the LaTeX exporter is able to control, this back-end is able to produce PDF output. Because the LaTeX exporter can be configured to use the hyperref package, the default setup produces fully-linked PDF output.

As in LaTeX, blank lines are meaningful for this back-end: a paragraph will not be started if two contiguous syntactical elements are not separated by an empty line.

This back-end also offers enhanced support for footnotes. Thus, it handles nested footnotes, footnotes in tables and footnotes in a list item's description.

Next: , Previous: LaTeX and PDF export, Up: LaTeX and PDF export

12.7.1 LaTeX export commands

C-c C-e l l     (org-latex-export-to-latex)
Export as a LaTeX file. For an Org file, the LaTeX file will be myfile.tex. The file will be overwritten without warning.
C-c C-e l L     (org-latex-export-as-latex)
Export to a temporary buffer. Do not create a file.
C-c C-e l p     (org-latex-export-to-pdf)
Export as LaTeX and then process to PDF.
C-c C-e l o
Export as LaTeX and then process to PDF, then open the resulting PDF file.

Next: , Previous: LaTeX export commands, Up: LaTeX and PDF export

12.7.2 Header and sectioning structure

By default, the first three outline levels become headlines, defining a general document structure. Additional levels are exported as itemize or enumerate lists. The transition can also occur at a different level (see Export settings).

By default, the LaTeX output uses the class article.

You can change this globally by setting a different value for org-latex-default-class or locally by adding an option like #+LATEX_CLASS: myclass in your file, or with a EXPORT_LATEX_CLASS property that applies when exporting a region containing only this (sub)tree. The class must be listed in org-latex-classes. This variable defines a header template for each class143, and allows you to define the sectioning structure for each class. You can also define your own classes there.

The LATEX_CLASS_OPTIONS keyword or EXPORT_LATEX_CLASS_OPTIONS property can specify the options for the \documentclass macro. These options have to be provided, as expected by LaTeX, within square brackets.

You can also use the LATEX_HEADER and LATEX_HEADER_EXTRA144 keywords in order to add lines to the header. See the docstring of org-latex-classes for more information.

An example is shown below.

     #+LATEX_CLASS: article
     #+LATEX_CLASS_OPTIONS: [a4paper]
     #+LATEX_HEADER: \usepackage{xyz}
     * Headline 1
       some text

Next: , Previous: Header and sectioning, Up: LaTeX and PDF export

12.7.3 Quoting LaTeX code

Embedded LaTeX as described in Embedded LaTeX, will be correctly inserted into the LaTeX file. Furthermore, you can add special code that should only be present in LaTeX export with the following constructs:

     Code within @@latex:some code@@ a paragraph.
     #+LATEX: Literal LaTeX code for export
     All lines between these markers are exported literally

Previous: Quoting LaTeX code, Up: LaTeX and PDF export

12.7.4 LaTeX specific attributes

LaTeX understands attributes specified in an ATTR_LATEX line. They affect tables, images, plain lists, special blocks and source blocks.

Tables in LaTeX export

For LaTeX export of a table, you can specify a label and a caption (see Images and tables). You can also use attributes to control table layout and contents. Valid LaTeX attributes include:

Nature of table's contents. It can be set to table, math, inline-math or verbatim. In particular, when in math or inline-math mode, every cell is exported as-is, horizontal rules are ignored and the table will be wrapped in a math environment. Also, contiguous tables sharing the same math mode will be wrapped within the same environment. Default mode is determined in org-latex-default-table-mode.
Environment used for the table. It can be set to any LaTeX table environment, like tabularx145, longtable, array, tabu146, bmatrix... It defaults to org-latex-default-table-environment value.
#+CAPTION keyword is the simplest way to set a caption for a table (see Images and tables). If you need more advanced commands for that task, you can use :caption attribute instead. Its value should be raw LaTeX code. It has precedence over #+CAPTION.
Float environment for the table. Possible values are sidewaystable, multicolumn, t and nil. When unspecified, a table with a caption will have a table environment. Moreover, :placement attribute can specify the positioning of the float.
Set, respectively, the alignment string of the table, its font size and its width. They only apply on regular tables.
Boolean specific to the tabu and longtabu environments, and only takes effect when used in conjunction with the :width attribute. When :spread is non-nil, the table will be spread or shrunk by the value of :width.
They toggle, respectively, booktabs usage (assuming the package is properly loaded), table centering and removal of every horizontal rule but the first one (in a "table.el" table only). In particular, org-latex-tables-booktabs (respectively org-latex-tables-centered) activates the first (respectively second) attribute globally.
A string that will be inserted, respectively, before the table within the math environment, after the table within the math environment, and between the macro name and the contents of the table. The :math-arguments attribute is used for matrix macros that require more than one argument (e.g., qbordermatrix).

Thus, attributes can be used in a wide array of situations, like writing a table that will span over multiple pages, or a matrix product:

     #+ATTR_LATEX: :environment longtable :align l|lp{3cm}r|l
     | ..... | ..... |
     | ..... | ..... |
     #+ATTR_LATEX: :mode math :environment bmatrix :math-suffix \times
     | a | b |
     | c | d |
     #+ATTR_LATEX: :mode math :environment bmatrix
     | 1 | 2 |
     | 3 | 4 |

In the example below, LaTeX command \bicaption{HeadingA}{HeadingB} will set the caption.

     #+ATTR_LATEX: :caption \bicaption{HeadingA}{HeadingB}
     | ..... | ..... |
     | ..... | ..... |
Images in LaTeX export

Images that are linked to without a description part in the link, like ‘[[file:img.jpg]]’ or ‘[[./img.jpg]]’ will be inserted into the PDF output file resulting from LaTeX processing. Org will use an \includegraphics macro to insert the image147.

You can specify specify image width or height with, respectively, :width and :height attributes. It is also possible to add any other option with the :options attribute, as shown in the following example:

     #+ATTR_LATEX: :width 5cm :options angle=90

If you need a specific command for the caption, use :caption attribute. It will override standard #+CAPTION value, if any.

     #+ATTR_LATEX: :caption \bicaption{HeadingA}{HeadingB}

If you have specified a caption as described in Images and tables, the picture will be wrapped into a figure environment and thus become a floating element. You can also ask Org to export an image as a float without specifying caption by setting the :float attribute. You may also set it to:

To modify the placement option of any floating environment, set the placement attribute.
     #+ATTR_LATEX: :float wrap :width 0.38\textwidth :placement {r}{0.4\textwidth}

If the :comment-include attribute is set to a non-nil value, the LaTeX \includegraphics macro will be commented out.

Plain lists in LaTeX export

Plain lists accept two optional attributes: :environment and :options. The first one allows the use of a non-standard environment (e.g., ‘inparaenum’). The second one specifies additional arguments for that environment.

     #+ATTR_LATEX: :environment compactitem :options [$\circ$]
     - you need ``paralist'' package to reproduce this example.
Source blocks in LaTeX export

In addition to syntax defined in Literal examples, names and captions (see Images and tables), source blocks also accept a :float attribute. You may set it to:

     #+ATTR_LATEX: :float nil
     #+BEGIN_SRC emacs-lisp
     Code that may not fit in a single page.
Special blocks in LaTeX export

In LaTeX back-end, special blocks become environments of the same name. Value of :options attribute will be appended as-is to that environment's opening string. For example:

     We demonstrate how to solve the Syracuse problem.
     #+ATTR_LATEX: :options [Proof of important theorem]
     Therefore, any even number greater than 2 is the sum of two primes.


     We demonstrate how to solve the Syracuse problem.
     \begin{proof}[Proof of important theorem]
     Therefore, any even number greater than 2 is the sum of two primes.

If you need to insert a specific caption command, use :caption attribute. It will override standard #+CAPTION value, if any. For example:

     #+ATTR_LATEX: :caption \MyCaption{HeadingA}
Horizontal rules

Width and thickness of a given horizontal rule can be controlled with, respectively, :width and :thickness attributes:

     #+ATTR_LATEX: :width .6\textwidth :thickness 0.8pt

Next: , Previous: LaTeX and PDF export, Up: Exporting

12.8 Markdown export

md export back-end generates Markdown syntax148 for an Org mode buffer.

It is built over HTML back-end: any construct not supported by Markdown syntax (e.g., tables) will be controlled and translated by html back-end (see HTML export).

Markdown export commands

C-c C-e m m     (org-md-export-to-markdown)
Export as a text file written in Markdown syntax. For an Org file,, the resulting file will be The file will be overwritten without warning.
C-c C-e m M     (org-md-export-as-markdown)
Export to a temporary buffer. Do not create a file.
C-c C-e m o
Export as a text file with Markdown syntax, then open it.

Header and sectioning structure

Markdown export can generate both atx and setext types for headlines, according to org-md-headline-style. The former introduces a hard limit of two levels, whereas the latter pushes it to six. Headlines below that limit are exported as lists. You can also set a soft limit before that one (see Export settings).

Next: , Previous: Markdown export, Up: Exporting

12.9 OpenDocument Text export

Org mode149 supports export to OpenDocument Text (ODT) format. Documents created by this exporter use the OpenDocument-v1.2 specification150 and are compatible with LibreOffice 3.4.

Next: , Previous: OpenDocument Text export, Up: OpenDocument Text export

12.9.1 Pre-requisites for ODT export

The ODT exporter relies on the zip program to create the final output. Check the availability of this program before proceeding further.

Next: , Previous: Pre-requisites for ODT export, Up: OpenDocument Text export

12.9.2 ODT export commands

Exporting to ODT

C-c C-e o o     (org-odt-export-to-odt)
Export as OpenDocument Text file.

If org-odt-preferred-output-format is specified, automatically convert the exported file to that format. See Automatically exporting to other formats.

For an Org file, the ODT file will be myfile.odt. The file will be overwritten without warning. If there is an active region,151 only the region will be exported. If the selected region is a single tree,152 the tree head will become the document title. If the tree head entry has, or inherits, an EXPORT_FILE_NAME property, that name will be used for the export.

C-c C-e o O Export as an OpenDocument Text file and open the resulting file.

If org-odt-preferred-output-format is specified, open the converted file instead. See Automatically exporting to other formats.

Next: , Previous: ODT export commands, Up: OpenDocument Text export

12.9.3 Extending ODT export

The ODT exporter can interface with a variety of document converters and supports popular converters out of the box. As a result, you can use it to export to formats like ‘doc’ or convert a document from one format (say ‘csv’) to another format (say ‘ods’ or ‘xls’).

If you have a working installation of LibreOffice, a document converter is pre-configured for you and you can use it right away. If you would like to use unoconv as your preferred converter, customize the variable org-odt-convert-process to point to unoconv. You can also use your own favorite converter or tweak the default settings of the LibreOffice and ‘unoconv’ converters. See Configuring a document converter. Automatically exporting to other formats

Very often, you will find yourself exporting to ODT format, only to immediately save the exported document to other formats like ‘doc’, ‘docx’, ‘rtf’, ‘pdf’ etc. In such cases, you can specify your preferred output format by customizing the variable org-odt-preferred-output-format. This way, the export commands (see Exporting to ODT) can be extended to export to a format that is of immediate interest to you. Converting between document formats

There are many document converters in the wild which support conversion to and from various file formats, including, but not limited to the ODT format. LibreOffice converter, mentioned above, is one such converter. Once a converter is configured, you can interact with it using the following command.

M-x org-odt-convert RET
Convert an existing document from one format to another. With a prefix argument, also open the newly produced file.

Next: , Previous: Extending ODT export, Up: OpenDocument Text export

12.9.4 Applying custom styles

The ODT exporter ships with a set of OpenDocument styles (see Working with OpenDocument style files) that ensure a well-formatted output. These factory styles, however, may not cater to your specific tastes. To customize the output, you can either modify the above styles files directly, or generate the required styles using an application like LibreOffice. The latter method is suitable for expert and non-expert users alike, and is described here. Applying custom styles: the easy way
  1. Create a sample file with the below settings and export it to ODT format.
              #+OPTIONS: H:10 num:t
  2. Open the above example.odt using LibreOffice. Use the Stylist to locate the target styles—these typically have the ‘Org’ prefix—and modify those to your taste. Save the modified file either as an OpenDocument Text (.odt) or OpenDocument Template (.ott) file.
  3. Customize the variable org-odt-styles-file and point it to the newly created file. For additional configuration options see Overriding factory styles.

    If you would like to choose a style on a per-file basis, you can use the #+ODT_STYLES_FILE option. A typical setting will look like

              #+ODT_STYLES_FILE: "/path/to/"


              #+ODT_STYLES_FILE: ("/path/to/" ("styles.xml" "image/hdr.png")) Using third-party styles and templates

You can use third-party styles and templates for customizing your output. This will produce the desired output only if the template provides all style names that the ‘ODT’ exporter relies on. Unless this condition is met, the output is going to be less than satisfactory. So it is highly recommended that you only work with templates that are directly derived from the factory settings.

Next: , Previous: Applying custom styles, Up: OpenDocument Text export

12.9.5 Links in ODT export

ODT exporter creates native cross-references for internal links. It creates Internet-style links for all other links.

A link with no description and destined to a regular (un-itemized) outline heading is replaced with a cross-reference and section number of the heading.

A ‘\ref{label}’-style reference to an image, table etc. is replaced with a cross-reference and sequence number of the labeled entity. See Labels and captions in ODT export.

Next: , Previous: Links in ODT export, Up: OpenDocument Text export

12.9.6 Tables in ODT export

Export of native Org mode tables (see Tables) and simple table.el tables is supported. However, export of complex table.el tables—tables that have column or row spans—is not supported. Such tables are stripped from the exported document.

By default, a table is exported with top and bottom frames and with rules separating row and column groups (see Column groups). Furthermore, all tables are typeset to occupy the same width. If the table specifies alignment and relative width for its columns (see Column width and alignment) then these are honored on export.153

You can control the width of the table by specifying :rel-width property using an #+ATTR_ODT line.

For example, consider the following table which makes use of all the rules mentioned above.

     #+ATTR_ODT: :rel-width 50
     | Area/Month    |   Jan |   Feb |   Mar |   Sum |
     | /             |     < |       |       |     < |
     | <l13>         |  <r5> |  <r5> |  <r5> |  <r6> |
     | North America |     1 |    21 |   926 |   948 |
     | Middle East   |     6 |    75 |   844 |   925 |
     | Asia Pacific  |     9 |    27 |   790 |   826 |
     | Sum           |    16 |   123 |  2560 |  2699 |

On export, the table will occupy 50% of text area. The columns will be sized (roughly) in the ratio of 13:5:5:5:6. The first column will be left-aligned and rest of the columns will be right-aligned. There will be vertical rules after separating the header and last columns from other columns. There will be horizontal rules separating the header and last rows from other rows.

If you are not satisfied with the above formatting options, you can create custom table styles and associate them with a table using the #+ATTR_ODT line. See Customizing tables in ODT export.

Next: , Previous: Tables in ODT export, Up: OpenDocument Text export

12.9.7 Images in ODT export

Embedding images

You can embed images within the exported document by providing a link to the desired image file with no link description. For example, to embed ‘img.png’ do either of the following:

Embedding clickable images

You can create clickable images by providing a link whose description is a link to an image file. For example, to embed a image org-mode-unicorn.png which when clicked jumps to website, do the following

Sizing and scaling of embedded images

You can control the size and scale of the embedded images using the #+ATTR_ODT attribute.

The exporter specifies the desired size of the image in the final document in units of centimeters. In order to scale the embedded images, the exporter queries for pixel dimensions of the images using one of a) ImageMagick's identify program or b) Emacs create-image and image-size APIs154. The pixel dimensions are subsequently converted in to units of centimeters using org-odt-pixels-per-inch. The default value of this variable is set to display-pixels-per-inch. You can tweak this variable to achieve the best results.

The examples below illustrate the various possibilities.

Explicitly size the image
To embed img.png as a 10 cm x 10 cm image, do the following:
          #+ATTR_ODT: :width 10 :height 10

Scale the image
To embed img.png at half its size, do the following:
          #+ATTR_ODT: :scale 0.5

Scale the image to a specific width
To embed img.png with a width of 10 cm while retaining the original height:width ratio, do the following:
          #+ATTR_ODT: :width 10

Scale the image to a specific height
To embed img.png with a height of 10 cm while retaining the original height:width ratio, do the following
          #+ATTR_ODT: :height 10
Anchoring of images

You can control the manner in which an image is anchored by setting the :anchor property of it's #+ATTR_ODT line. You can specify one of the following three values for the :anchor property: ‘"as-char"’, ‘"paragraph"’ and ‘"page"’.

To create an image that is anchored to a page, do the following:

     #+ATTR_ODT: :anchor "page"

Next: , Previous: Images in ODT export, Up: OpenDocument Text export

12.9.8 Math formatting in ODT export

The ODT exporter has special support for handling math.

Next: , Previous: Math formatting in ODT export, Up: Math formatting in ODT export Working with LaTeX math snippets

LaTeX math snippets (see LaTeX fragments) can be embedded in the ODT document in one of the following ways:

  1. MathML

    This option is activated on a per-file basis with

              #+OPTIONS: LaTeX:t

    With this option, LaTeX fragments are first converted into MathML fragments using an external LaTeX-to-MathML converter program. The resulting MathML fragments are then embedded as an OpenDocument Formula in the exported document.

    You can specify the LaTeX-to-MathML converter by customizing the variables org-latex-to-mathml-convert-command and org-latex-to-mathml-jar-file.

    If you prefer to use MathToWeb155 as your converter, you can configure the above variables as shown below.

              (setq org-latex-to-mathml-convert-command
                    "java -jar %j -unicode -force -df %o %I"

    You can use the following commands to quickly verify the reliability of the LaTeX-to-MathML converter.

    M-x org-odt-export-as-odf RET
    Convert a LaTeX math snippet to an OpenDocument formula (.odf) file.
    M-x org-odt-export-as-odf-and-open RET
    Convert a LaTeX math snippet to an OpenDocument formula (.odf) file and open the formula file with the system-registered application.

  2. PNG images

    This option is activated on a per-file basis with

              #+OPTIONS: tex:dvipng


              #+OPTIONS: tex:imagemagick

    With this option, LaTeX fragments are processed into PNG images and the resulting images are embedded in the exported document. This method requires that the dvipng program or imagemagick suite be available on your system.

Previous: Working with LaTeX math snippets, Up: Math formatting in ODT export Working with MathML or OpenDocument formula files

For various reasons, you may find embedding LaTeX math snippets in an ODT document less than reliable. In that case, you can embed a math equation by linking to its MathML (.mml) source or its OpenDocument formula (.odf) file as shown below:




Next: , Previous: Math formatting in ODT export, Up: OpenDocument Text export

12.9.9 Labels and captions in ODT export

You can label and caption various category of objects—an inline image, a table, a LaTeX fragment or a Math formula—using #+LABEL and #+CAPTION lines. See Images and tables. ODT exporter enumerates each labeled or captioned object of a given category separately. As a result, each such object is assigned a sequence number based on order of it's appearance in the Org file.

In the exported document, a user-provided caption is augmented with the category and sequence number. Consider the following inline image in an Org file.

     #+CAPTION: Bell curve
     #+LABEL:   fig:SED-HR4049

It could be rendered as shown below in the exported document.

     Figure 2: Bell curve

You can modify the category component of the caption by customizing the option org-odt-category-map-alist. For example, to tag all embedded images with the string ‘Illustration’ (instead of the default ‘Figure’) use the following setting:

     (setq org-odt-category-map-alist
           (("__Figure__" "Illustration" "value" "Figure" org-odt--enumerable-image-p)))

With this, previous image will be captioned as below in the exported document.

     Illustration 2: Bell curve

Next: , Previous: Labels and captions in ODT export, Up: OpenDocument Text export

12.9.10 Literal examples in ODT export

Export of literal examples (see Literal examples) with full fontification is supported. Internally, the exporter relies on htmlfontify.el to generate all style definitions needed for a fancy listing.156 The auto-generated styles have ‘OrgSrc’ as prefix and inherit their color from the faces used by Emacs font-lock library for the source language.

If you prefer to use your own custom styles for fontification, you can do so by customizing the option org-odt-create-custom-styles-for-srcblocks.

You can turn off fontification of literal examples by customizing the option org-odt-fontify-srcblocks.

Previous: Literal examples in ODT export, Up: OpenDocument Text export

12.9.11 Advanced topics in ODT export

If you rely heavily on ODT export, you may want to exploit the full set of features that the exporter offers. This section describes features that would be of interest to power users.

Next: , Previous: Advanced topics in ODT export, Up: Advanced topics in ODT export Configuring a document converter

The ODT exporter can work with popular converters with little or no extra configuration from your side. See Extending ODT export. If you are using a converter that is not supported by default or if you would like to tweak the default converter settings, proceed as below.

  1. Register the converter

    Name your converter and add it to the list of known converters by customizing the option org-odt-convert-processes. Also specify how the converter can be invoked via command-line to effect the conversion.

  2. Configure its capabilities

    Specify the set of formats the converter can handle by customizing the variable org-odt-convert-capabilities. Use the default value for this variable as a guide for configuring your converter. As suggested by the default setting, you can specify the full set of formats supported by the converter and not limit yourself to specifying formats that are related to just the OpenDocument Text format.

  3. Choose the converter

    Select the newly added converter as the preferred one by customizing the option org-odt-convert-process.

Next: , Previous: Configuring a document converter, Up: Advanced topics in ODT export Working with OpenDocument style files

This section explores the internals of the ODT exporter and the means by which it produces styled documents. Read this section if you are interested in exploring the automatic and custom OpenDocument styles used by the exporter.

Factory styles

The ODT exporter relies on two files for generating its output. These files are bundled with the distribution under the directory pointed to by the variable org-odt-styles-dir. The two files are:

Overriding factory styles

The following two variables control the location from which the ODT exporter picks up the custom styles and content template files. You can customize these variables to override the factory styles used by the exporter.

Next: , Previous: Working with OpenDocument style files, Up: Advanced topics in ODT export Creating one-off styles

There are times when you would want one-off formatting in the exported document. You can achieve this by embedding raw OpenDocument XML in the Org file. The use of this feature is better illustrated with couple of examples.

  1. Embedding ODT tags as part of regular text

    You can inline OpenDocument syntax by enclosing it within ‘@@odt:...@@’ markup. For example, to highlight a region of text do the following:

              @@odt:<text:span text:style-name="Highlight">This is a highlighted
              text</text:span>@@.  But this is a regular text.

    Hint: To see the above example in action, edit your styles.xml (see Factory styles) and add a custom ‘Highlight’ style as shown below.

              <style:style style:name="Highlight" style:family="text">
                <style:text-properties fo:background-color="#ff0000"/>
  2. Embedding a one-line OpenDocument XML

    You can add a simple OpenDocument one-liner using the #+ODT: directive. For example, to force a page break do the following:

              #+ODT: <text:p text:style-name="PageBreak"/>

    Hint: To see the above example in action, edit your styles.xml (see Factory styles) and add a custom ‘PageBreak’ style as shown below.

              <style:style style:name="PageBreak" style:family="paragraph"
                <style:paragraph-properties fo:break-before="page"/>
  3. Embedding a block of OpenDocument XML

    You can add a large block of OpenDocument XML using the #+BEGIN_ODT...#+END_ODT construct.

    For example, to create a one-off paragraph that uses bold text, do the following:

              <text:p text:style-name="Text_20_body_20_bold">
              This paragraph is specially formatted and uses bold text.

Next: , Previous: Creating one-off styles, Up: Advanced topics in ODT export Customizing tables in ODT export

You can override the default formatting of the table by specifying a custom table style with the #+ATTR_ODT line. For a discussion on default formatting of tables see Tables in ODT export.

This feature closely mimics the way table templates are defined in the OpenDocument-v1.2 specification.157

Custom table styles: an illustration

To have a quick preview of this feature, install the below setting and export the table that follows:

     (setq org-odt-table-styles
           (append org-odt-table-styles
                 '(("TableWithHeaderRowAndColumn" "Custom"
                     ((use-first-row-styles . t)
                      (use-first-column-styles . t)))
                     ("TableWithFirstRowandLastRow" "Custom"
                      ((use-first-row-styles . t)
                      (use-last-row-styles . t))))))
     #+ATTR_ODT: :style "TableWithHeaderRowAndColumn"
     | Name  | Phone | Age |
     | Peter |  1234 |  17 |
     | Anna  |  4321 |  25 |

In the above example, you used a template named ‘Custom’ and installed two table styles with the names ‘TableWithHeaderRowAndColumn’ and ‘TableWithFirstRowandLastRow’. (Important: The OpenDocument styles needed for producing the above template have been pre-defined for you. These styles are available under the section marked ‘Custom Table Template’ in OrgOdtContentTemplate.xml (see Factory styles). If you need additional templates you have to define these styles yourselves.

Custom table styles: the nitty-gritty

To use this feature proceed as follows:

  1. Create a table template158

    A table template is nothing but a set of ‘table-cell’ and ‘paragraph’ styles for each of the following table cell categories:

    • Body
    • First column
    • Last column
    • First row
    • Last row
    • Even row
    • Odd row
    • Even column
    • Odd Column

    The names for the above styles must be chosen based on the name of the table template using a well-defined convention.

    The naming convention is better illustrated with an example. For a table template with the name ‘Custom’, the needed style names are listed in the following table.

    Table cell type table-cell style paragraph style

    Body CustomTableCellCustomTableParagraph
    First column CustomFirstColumnTableCellCustomFirstColumnTableParagraph
    Last column CustomLastColumnTableCellCustomLastColumnTableParagraph
    First row CustomFirstRowTableCellCustomFirstRowTableParagraph
    Last row CustomLastRowTableCellCustomLastRowTableParagraph
    Even row CustomEvenRowTableCellCustomEvenRowTableParagraph
    Odd row CustomOddRowTableCellCustomOddRowTableParagraph
    Even column CustomEvenColumnTableCellCustomEvenColumnTableParagraph
    Odd column CustomOddColumnTableCellCustomOddColumnTableParagraph

    To create a table template with the name ‘Custom’, define the above styles in the <office:automatic-styles>...</office:automatic-styles> element of the content template file (see Factory styles).

  2. Define a table style159

    To define a table style, create an entry for the style in the variable org-odt-table-styles and specify the following:

    • the name of the table template created in step (1)
    • the set of cell styles in that template that are to be activated

    For example, the entry below defines two different table styles ‘TableWithHeaderRowAndColumn’ and ‘TableWithFirstRowandLastRow’ based on the same template ‘Custom’. The styles achieve their intended effect by selectively activating the individual cell styles in that template.

              (setq org-odt-table-styles
                    (append org-odt-table-styles
                            '(("TableWithHeaderRowAndColumn" "Custom"
                               ((use-first-row-styles . t)
                                (use-first-column-styles . t)))
                              ("TableWithFirstRowandLastRow" "Custom"
                               ((use-first-row-styles . t)
                                (use-last-row-styles . t))))))
  3. Associate a table with the table style

    To do this, specify the table style created in step (2) as part of the ATTR_ODT line as shown below.

              #+ATTR_ODT: :style "TableWithHeaderRowAndColumn"
              | Name  | Phone | Age |
              | Peter |  1234 |  17 |
              | Anna  |  4321 |  25 |

Previous: Customizing tables in ODT export, Up: Advanced topics in ODT export Validating OpenDocument XML

Occasionally, you will discover that the document created by the ODT exporter cannot be opened by your favorite application. One of the common reasons for this is that the .odt file is corrupt. In such cases, you may want to validate the document against the OpenDocument RELAX NG Compact Syntax (RNC) schema.

For de-compressing the .odt file160: see File Archives. For general help with validation (and schema-sensitive editing) of XML files: see Introduction.

If you have ready access to OpenDocument .rnc files and the needed schema-locating rules in a single folder, you can customize the variable org-odt-schema-dir to point to that directory. The ODT exporter will take care of updating the rng-schema-locating-files for you.

Next: , Previous: OpenDocument Text export, Up: Exporting

12.10 Org export

org export back-end creates a normalized version of the Org document in current buffer. In particular, it evaluates Babel code (see Evaluating code blocks) and removes other back-ends specific contents.

Org export commands

C-c C-e O o     (org-org-export-to-org)
Export as an Org document. For an Org file,, the resulting file will be The file will be overwritten without warning.
C-c C-e O O     (org-org-export-as-org)
Export to a temporary buffer. Do not create a file.
C-c C-e O v
Export to an Org file, then open it.

Next: , Previous: Org export, Up: Exporting

12.11 Texinfo export

texinfo’ export back-end generates Texinfo code and can compile it into an Info file.

Next: , Previous: Texinfo export, Up: Texinfo export

12.11.1 Texinfo export commands

C-c C-e i t     (org-texinfo-export-to-texinfo)
Export as a Texinfo file. For an Org file,, the resulting file will be myfile.texi. The file will be overwritten without warning.
C-c C-e i i     (org-texinfo-export-to-info)
Export to Texinfo and then process to an Info file161.

Next: , Previous: Texinfo export commands, Up: Texinfo export

12.11.2 Document preamble

When processing a document, ‘texinfo’ back-end generates a minimal file header along with a title page, a copyright page, and a menu. You control the latter through the structure of the document (see Headings and sectioning structure). Various keywords allow you to tweak the other parts. It is also possible to give directions to install the document in the ‘Top’ node.

File header

Upon creating the header of a Texinfo file, the back-end guesses a name for the Info file to be compiled. This may not be a sensible choice, e.g., if you want to produce the final document in a different directory. Specify an alternate path with #+TEXINFO_FILENAME keyword to override the default destination.

Along with the output file name, the header contains information about the language (see Export settings) and current encoding used162. Insert a #+TEXINFO_HEADER keyword for each additional command needed, e.g., @code{@synindex}.

If you happen to regularly install the same set of commands, it may be easier to define your own class in org-texinfo-classes, which see. Set #+TEXINFO_CLASS keyword accordingly in your document to activate it.

Title and copyright page

The default template includes a title page for hard copy output. The title and author displayed on this page are extracted from, respectively, #+TITLE and #+AUTHOR keywords (see Export settings). It is also possible to print a different, more specific, title with #+TEXINFO_PRINTED_TITLE keyword, and add subtitles with #+SUBTITLE keyword. Both expect raw Texinfo code in their value.

Likewise, information brought by #+AUTHOR may not be enough. You can include other authors with several #+SUBAUTHOR keywords. Values are also expected to be written in Texinfo code.

     #+AUTHOR: Jane Smith
     #+SUBAUTHOR: John Doe
     #+TEXINFO_PRINTED_TITLE: This Long Title@inlinefmt{tex,@*} Is Broken in @TeX{}

Copying material is defined in a dedicated headline with a non-nil :COPYING: property. The contents are inserted within a @copying command at the beginning of the document whereas the heading itself does not appear in the structure of the document.

Copyright information is printed on the back of the title page.

     * Copying
       :COPYING: t
       This is a short example of a complete Texinfo file, version 1.0.
       Copyright \copy 2017 Free Software Foundation, Inc.
The Top node

You may ultimately want to install your new Info file to your system. You can write an appropriate entry in the top level directory specifying its category and title with, respectively, #+TEXINFO_DIR_CATEGORY and #+TEXINFO_DIR_TITLE. Optionally, you can add a short description using #+TEXINFO_DIR_DESC. The following example would write an entry similar to Org's in the ‘Top’ node.

     #+TEXINFO_DIR_TITLE: Org Mode: (org)
     #+TEXINFO_DIR_DESC: Outline-based notes management and organizer

Next: , Previous: Document preamble, Up: Texinfo export

12.11.3 Headings and sectioning structure

texinfo’ uses a pre-defined scheme, or class, to convert headlines into Texinfo structuring commands. For example, a top level headline appears as @chapter if it should be numbered or as @unnumbered otherwise. If you need to use a different set of commands, e.g., to start with @part instead of @chapter, install a new class in org-texinfo-classes, then activate it with #+TEXINFO_CLASS keyword. Export process defaults to org-texinfo-default-class when there is no such keyword in the document.

If a headline's level has no associated structuring command, or is below a certain threshold see Export settings, that headline becomes a list in Texinfo output.

As an exception, a headline with a non-nil :APPENDIX: property becomes an appendix, independently on its level and the class used.

Each regular sectioning structure creates a menu entry, named after the heading. You can provide a different, e.g., shorter, title in :ALT_TITLE: property (see Table of contents). Optionally, you can specify a description for the item in :DESCRIPTION: property. E.g.,

     * Controlling Screen Display
       :ALT_TITLE: Display
       :DESCRIPTION: Controlling Screen Display

Next: , Previous: Headings and sectioning structure, Up: Texinfo export

12.11.4 Indices

Index entries are created using dedicated keywords. ‘texinfo’ back-end provides one for each predefined type: #+CINDEX, #+FINDEX, #+KINDEX, #+PINDEX, #+TINDEX and #+VINDEX. For custom indices, you can write raw Texinfo code (see Quoting Texinfo code).

     #+CINDEX: Defining indexing entries

To generate an index, you need to set the :INDEX: property of a headline to an appropriate abbreviation (e.g., ‘cp’ or ‘vr’). The headline is then exported as an unnumbered chapter or section command and the index is inserted after its contents.

     * Concept Index
       :INDEX: cp

Next: , Previous: Indices, Up: Texinfo export

12.11.5 Quoting Texinfo code

It is possible to insert raw Texinfo code using any of the following constructs

     Richard @@texinfo:@sc{@@Stallman@@texinfo:}@@ commence' GNU.
     #+TEXINFO: @need800
     This paragraph is preceded by...
     @auindex Johnson, Mark
     @auindex Lakoff, George

Next: , Previous: Quoting Texinfo code, Up: Texinfo export

12.11.6 Texinfo specific attributes

texinfo’ back-end understands several attributes in plain lists and tables. They must be specified using an #+ATTR_TEXINFO keyword, written just above the list or table.

Plain lists

In Texinfo output, description lists appear as two-column tables, using the default command @table. You can use @ftable or @vtable163 instead with :table-type attribute.

In any case, these constructs require a highlighting command for entries in the list. You can provide one with :indic attribute. If you do not, it defaults to the value stored in org-texinfo-def-table-markup, which see.

     #+ATTR_TEXINFO: :indic @asis
     - foo :: This is the text for /foo/, with no highlighting.

When exporting a table, column widths are deduced from the longest cell in each column. You can also define them explicitly as fractions of the line length, using :columns attribute.

     #+ATTR_TEXINFO: :columns .5 .5
     | a cell | another cell |

Previous: Texinfo specific attributes, Up: Texinfo export

12.11.7 An example

Here is a thorough example, taken from see GNU Sample Texts.

     #+MACRO: version 2.0
     #+MACRO: updated last updated 4 March 2014
     #+OPTIONS: ':t toc:t author:t email:t
     #+TITLE: GNU Sample {{{version}}}
     #+AUTHOR: A.U. Thor
     #+LANGUAGE: en
     #+TEXINFO_HEADER: @syncodeindex pg cp
     #+TEXINFO_DIR_CATEGORY: Texinfo documentation system
     #+TEXINFO_DIR_TITLE: sample: (sample)
     #+TEXINFO_DIR_DESC: Invoking sample
     #+SUBTITLE: for version 2.0, last updated 4 March 2014
     * Copying
       :COPYING:  t
       This manual is for GNU Sample (version {{{version}}},
       {{{updated}}}), which is an example in the Texinfo documentation.
       Copyright @@texinfo:@copyright{}@@ 2013 Free Software Foundation,
       Permission is granted to copy, distribute and/or modify this
       document under the terms of the GNU Free Documentation License,
       Version 1.3 or any later version published by the Free Software
       Foundation; with no Invariant Sections, with no Front-Cover Texts,
       and with no Back-Cover Texts.  A copy of the license is included in
       the section entitled "GNU Free Documentation License".
     * Invoking sample
       #+PINDEX: sample
       #+CINDEX: invoking @command{sample}
       This is a sample manual.  There is no sample program to invoke, but
       if there were, you could see its basic usage and command line
       options here.
     * GNU Free Documentation License
       :APPENDIX: t
       #+TEXINFO: @include fdl.texi
     * Index
       :INDEX:    cp

Next: , Previous: Texinfo export, Up: Exporting

12.12 iCalendar export

Some people use Org mode for keeping track of projects, but still prefer a standard calendar application for anniversaries and appointments. In this case it can be useful to show deadlines and other time-stamped items in Org files in the calendar application. Org mode can export calendar information in the standard iCalendar format. If you also want to have TODO entries included in the export, configure the variable org-icalendar-include-todo. Plain timestamps are exported as VEVENT, and TODO items as VTODO. It will also create events from deadlines that are in non-TODO items. Deadlines and scheduling dates in TODO items will be used to set the start and due dates for the TODO entry164. As categories, it will use the tags locally defined in the heading, and the file/tree category165. See the variable org-icalendar-alarm-time for a way to assign alarms to entries with a time.

The iCalendar standard requires each entry to have a globally unique identifier (UID). Org creates these identifiers during export. If you set the variable org-icalendar-store-UID, the UID will be stored in the :ID: property of the entry and re-used next time you report this entry. Since a single entry can give rise to multiple iCalendar entries (as a timestamp, a deadline, a scheduled item, and as a TODO item), Org adds prefixes to the UID, depending on what triggered the inclusion of the entry. In this way the UID remains unique, but a synchronization program can still figure out from which entry all the different instances originate.

C-c C-e c f     (org-icalendar-export-to-ics)
Create iCalendar entries for the current buffer and store them in the same directory, using a file extension .ics.
C-c C-e c a     (org-icalendar-export-agenda-files)
Like C-c C-e c f, but do this for all files in org-agenda-files. For each of these files, a separate iCalendar file will be written.
C-c C-e c c     (org-icalendar-combine-agenda-files)
Create a single large iCalendar file from all files in org-agenda-files and write it to the file given by org-icalendar-combined-agenda-file.

The export will honor SUMMARY, DESCRIPTION and LOCATION166 properties if the selected entries have them. If not, the summary will be derived from the headline, and the description from the body (limited to org-icalendar-include-body characters).

How this calendar is best read and updated, depends on the application you are using. The FAQ covers this issue.

Next: , Previous: iCalendar export, Up: Exporting

12.13 Other built-in back-ends

On top of the aforementioned back-ends, Org comes with other built-in ones:

To activate these export back-end, customize org-export-backends or load them directly with e.g., (require 'ox-man). This will add new keys in the export dispatcher (see The Export Dispatcher).

See the comment section of these files for more information on how to use them.

Next: , Previous: Other built-in back-ends, Up: Exporting

12.14 Export in foreign buffers

Most built-in back-ends come with a command to convert the selected region into a selected format and replace this region by the exported output. Here is a list of such conversion commands:

Convert the selected region into HTML.
Convert the selected region into LaTeX.
Convert the selected region into Texinfo.
Convert the selected region into MarkDown.

This is particularly useful for converting tables and lists in foreign buffers. E.g., in an HTML buffer, you can turn on orgstruct-mode, then use Org commands for editing a list, and finally select and convert the list with M-x org-html-convert-region-to-html RET.

Previous: Export in foreign buffers, Up: Exporting

12.15 Advanced configuration


Two hooks are run during the first steps of the export process. The first one, org-export-before-processing-hook is called before expanding macros, Babel code and include keywords in the buffer. The second one, org-export-before-parsing-hook, as its name suggests, happens just before parsing the buffer. Their main use is for heavy duties, that is duties involving structural modifications of the document. For example, one may want to remove every headline in the buffer during export. The following code can achieve this:

     (defun my-headline-removal (backend)
       "Remove all headlines in the current buffer.
     BACKEND is the export back-end being used, as a symbol."
        (lambda () (delete-region (point) (progn (forward-line) (point))))))
     (add-hook 'org-export-before-parsing-hook 'my-headline-removal)

Note that functions used in these hooks require a mandatory argument, a symbol representing the back-end used.


Filters are lists of functions applied on a specific part of the output from a given back-end. More explicitly, each time a back-end transforms an Org object or element into another language, all functions within a given filter type are called in turn on the string produced. The string returned by the last function will be the one used in the final output.

There are filters sets for each type of element or object, for plain text, for the parse tree, for the export options and for the final output. They are all named after the same scheme: org-export-filter-TYPE-functions, where TYPE is the type targeted by the filter. Valid types are:

bold babel-call center-block
clock code comment
comment-block diary-sexp drawer
dynamic-block entity example-block
export-block export-snippet final-output
fixed-width footnote-definition footnote-reference
headline horizontal-rule inline-babel-call
inline-src-block inlinetask italic
item keyword latex-environment
latex-fragment line-break link
node-property options paragraph
parse-tree plain-list plain-text
planning property-drawer quote-block
quote-section radio-target section
special-block src-block statistics-cookie
strike-through subscript superscript
table table-cell table-row
target timestamp underline
verbatim verse-block

For example, the following snippet allows me to use non-breaking spaces in the Org buffer and get them translated into LaTeX without using the \nbsp macro (where _ stands for the non-breaking space):

     (defun my-latex-filter-nobreaks (text backend info)
       "Ensure \"_\" are properly handled in LaTeX export."
       (when (org-export-derived-backend-p backend 'latex)
             (replace-regexp-in-string "_" "~" text)))
     (add-to-list 'org-export-filter-plain-text-functions

Three arguments must be provided to a filter: the code being changed, the back-end used, and some information about the export process. You can safely ignore the third argument for most purposes. Note the use of org-export-derived-backend-p, which ensures that the filter will only be applied when using latex back-end or any other back-end derived from it (e.g., beamer).

Extending an existing back-end

This is obviously the most powerful customization, since the changes happen at the parser level. Indeed, some export back-ends are built as extensions of other ones (e.g., Markdown back-end an extension of HTML back-end).

Extending a back-end means that if an element type is not transcoded by the new back-end, it will be handled by the original one. Hence you can extend specific parts of a back-end without too much work.

As an example, imagine we want the ascii back-end to display the language used in a source block, when it is available, but only when some attribute is non-nil, like the following:

     #+ATTR_ASCII: :language t

Because that back-end is lacking in that area, we are going to create a new back-end, my-ascii that will do the job.

     (defun my-ascii-src-block (src-block contents info)
       "Transcode a SRC-BLOCK element from Org to ASCII.
     CONTENTS is nil.  INFO is a plist used as a communication
       (if (not (org-export-read-attribute :attr_ascii src-block :language))
         (org-export-with-backend 'ascii src-block contents info)
        (format ",--[ %s ]--\n%s`----"
                (org-element-property :language src-block)
                 "^" "| "
                  (org-export-format-code-default src-block info)))))))
     (org-export-define-derived-backend 'my-ascii 'ascii
       :translate-alist '((src-block . my-ascii-src-block)))

The my-ascii-src-block function looks at the attribute above the element. If it isn't true, it gives hand to the ascii back-end. Otherwise, it creates a box around the code, leaving room for the language. A new back-end is then created. It only changes its behavior when translating src-block type element. Now, all it takes to use the new back-end is calling the following from an Org buffer:

     (org-export-to-buffer 'my-ascii "*Org MY-ASCII Export*")

It is obviously possible to write an interactive function for this, install it in the export dispatcher menu, and so on.

Next: , Previous: Exporting, Up: Top

13 Publishing

Org includes a publishing management system that allows you to configure automatic HTML conversion of projects composed of interlinked org files. You can also configure Org to automatically upload your exported HTML pages and related attachments, such as images and source code files, to a web server.

You can also use Org to convert files into PDF, or even combine HTML and PDF conversion so that files are available in both formats on the server.

Publishing has been contributed to Org by David O'Toole.

Next: , Previous: Publishing, Up: Publishing

13.1 Configuration

Publishing needs significant configuration to specify files, destination and many other properties of a project.

Next: , Previous: Configuration, Up: Configuration

13.1.1 The variable org-publish-project-alist

Publishing is configured almost entirely through setting the value of one variable, called org-publish-project-alist. Each element of the list configures one project, and may be in one of the two following forms:

        ("project-name" :property value :property value ...)
          i.e., a well-formed property list with alternating keys and values
        ("project-name" :components ("project-name" "project-name" ...))

In both cases, projects are configured by specifying property values. A project defines the set of files that will be published, as well as the publishing configuration to use when publishing those files. When a project takes the second form listed above, the individual members of the :components property are taken to be sub-projects, which group together files requiring different publishing options. When you publish such a “meta-project”, all the components will also be published, in the sequence given.

Next: , Previous: Project alist, Up: Configuration

13.1.2 Sources and destinations for files

Most properties are optional, but some should always be set. In particular, Org needs to know where to look for source files, and where to put published files.

:base-directory Directory containing publishing source files
:publishing-directory Directory where output files will be published. You can directly publish to a web server using a file name syntax appropriate for the Emacs tramp package. Or you can publish to a local directory and use external tools to upload your website (see Uploading files).
:preparation-function Function or list of functions to be called before starting the publishing process, for example, to run make for updating files to be published. The project property list is scoped into this call as the variable project-plist.
:completion-function Function or list of functions called after finishing the publishing process, for example, to change permissions of the resulting files. The project property list is scoped into this call as the variable project-plist.

Next: , Previous: Sources and destinations, Up: Configuration

13.1.3 Selecting files

By default, all files with extension .org in the base directory are considered part of the project. This can be modified by setting the properties

:base-extension Extension (without the dot!) of source files. This actually is a regular expression. Set this to the symbol any if you want to get all files in :base-directory, even without extension.

:exclude Regular expression to match file names that should not be published, even though they have been selected on the basis of their extension.

:include List of files to be included regardless of :base-extension and :exclude.

:recursive non-nil means, check base-directory recursively for files to publish.

Next: , Previous: Selecting files, Up: Configuration

13.1.4 Publishing action

Publishing means that a file is copied to the destination directory and possibly transformed in the process. The default transformation is to export Org files as HTML files, and this is done by the function org-html-publish-to-html, which calls the HTML exporter (see HTML export). But you also can publish your content as PDF files using org-latex-publish-to-pdf or as ascii, Texinfo, etc., using the corresponding functions.

If you want to publish the Org file as an .org file but with the archived, commented and tag-excluded trees removed, use the function org-org-publish-to-org. This will produce and put it in the publishing directory. If you want a htmlized version of this file, set the parameter :htmlized-source to t, it will produce in the publishing directory167.

Other files like images only need to be copied to the publishing destination. For this you can use org-publish-attachment. For non-org files, you always need to specify the publishing function:

:publishing-function Function executing the publication of a file. This may also be a list of functions, which will all be called in turn.
:htmlized-source non-nil means, publish htmlized source.

The function must accept three arguments: a property list containing at least a :publishing-directory property, the name of the file to be published and the path to the publishing directory of the output file. It should take the specified file, make the necessary transformation (if any) and place the result into the destination folder.

Next: , Previous: Publishing action, Up: Configuration

13.1.5 Options for the exporters

The property list can be used to set many export options for the exporters. In most cases, these properties correspond to user variables in Org. The first table below lists these properties along with the variable they belong to. The second table list HTML specific properties. See the documentation string of these options for details.

:archived-trees org-export-with-archived-trees
:exclude-tags org-export-exclude-tags
:headline-levels org-export-headline-levels
:language org-export-default-language
:preserve-breaks org-export-preserve-breaks
:section-numbers org-export-with-section-numbers
:select-tags org-export-select-tags
:with-author org-export-with-author
:with-creator org-export-with-creator
:with-drawers org-export-with-drawers
:with-email org-export-with-email
:with-emphasize org-export-with-emphasize
:with-fixed-width org-export-with-fixed-width
:with-footnotes org-export-with-footnotes
:with-latex org-export-with-latex
:with-planning org-export-with-planning
:with-priority org-export-with-priority
:with-special-strings org-export-with-special-strings
:with-sub-superscript org-export-with-sub-superscripts
:with-tables org-export-with-tables
:with-tags org-export-with-tags
:with-tasks org-export-with-tasks
:with-timestamps org-export-with-timestamps
:with-toc org-export-with-toc
:with-todo-keywords org-export-with-todo-keywords

:html-doctype org-html-doctype
:html-container org-html-container-element
:html-html5-fancy org-html-html5-fancy
:html-xml-declaration org-html-xml-declaration
:html-link-up org-html-link-up
:html-link-home org-html-link-home
:html-link-org-as-html org-html-link-org-files-as-html
:html-link-use-abs-url org-html-link-use-abs-url
:html-head org-html-head
:html-head-extra org-html-head-extra
:html-inline-images org-html-inline-images
:html-extension org-html-extension
:html-preamble org-html-preamble
:html-postamble org-html-postamble
:html-table-attributes org-html-table-default-attributes
:html-table-row-tags org-html-table-row-tags
:html-head-include-default-style org-html-head-include-default-style
:html-head-include-scripts org-html-head-include-scripts

Most of the org-export-with-* variables have the same effect in each exporter.

When a property is given a value in org-publish-project-alist, its setting overrides the value of the corresponding user variable (if any) during publishing. Options set within a file (see Export settings), however, override everything.

Next: , Previous: Publishing options, Up: Configuration

13.1.6 Links between published files

To create a link from one Org file to another, you would use something like ‘[[][The foo]]’ or simply ‘’ (see Hyperlinks). When published, this link becomes a link to foo.html. You can thus interlink the pages of your "org web" project and the links will work as expected when you publish them to HTML. If you also publish the Org source file and want to link to it, use an http: link instead of a file: link, because file: links are converted to link to the corresponding html file.

You may also link to related files, such as images. Provided you are careful with relative file names, and provided you have also configured Org to upload the related files, these links will work too. See Complex example, for an example of this usage.

Next: , Previous: Publishing links, Up: Configuration

13.1.7 Generating a sitemap

The following properties may be used to control publishing of a map of files for a given project.

:auto-sitemap When non-nil, publish a sitemap during org-publish-current-project or org-publish-all.

:sitemap-filename Filename for output of sitemap. Defaults to (which becomes sitemap.html).

:sitemap-title Title of sitemap page. Defaults to name of file.

:sitemap-function Plug-in function to use for generation of the sitemap. Defaults to org-publish-org-sitemap, which generates a plain list of links to all files in the project.

:sitemap-sort-folders Where folders should appear in the sitemap. Set this to first (default) or last to display folders first or last, respectively. Any other value will mix files and folders.

:sitemap-sort-files How the files are sorted in the site map. Set this to alphabetically (default), chronologically or anti-chronologically. chronologically sorts the files with older date first while anti-chronologically sorts the files with newer date first. alphabetically sorts the files alphabetically. The date of a file is retrieved with org-publish-find-date.

:sitemap-ignore-case Should sorting be case-sensitive? Default nil.

:sitemap-file-entry-format With this option one can tell how a sitemap's entry is formatted in the sitemap. This is a format string with some escape sequences: %t stands for the title of the file, %a stands for the author of the file and %d stands for the date of the file. The date is retrieved with the org-publish-find-date function and formatted with org-publish-sitemap-date-format. Default %t.

:sitemap-date-format Format string for the format-time-string function that tells how a sitemap entry's date is to be formatted. This property bypasses org-publish-sitemap-date-format which defaults to %Y-%m-%d.

:sitemap-sans-extension When non-nil, remove filenames' extensions from the generated sitemap. Useful to have cool URIs (see Defaults to nil.

Previous: Sitemap, Up: Configuration

13.1.8 Generating an index

Org mode can generate an index across the files of a publishing project.

:makeindex When non-nil, generate in index in the file and publish it as theindex.html.

The file will be created when first publishing a project with the :makeindex set. The file only contains a statement #+INCLUDE: "". You can then build around this include statement by adding a title, style information, etc.

Next: , Previous: Configuration, Up: Publishing

13.2 Uploading files

For those people already utilizing third party sync tools such as rsync or unison, it might be preferable not to use the built in remote publishing facilities of Org mode which rely heavily on Tramp. Tramp, while very useful and powerful, tends not to be so efficient for multiple file transfer and has been known to cause problems under heavy usage.

Specialized synchronization utilities offer several advantages. In addition to timestamp comparison, they also do content and permissions/attribute checks. For this reason you might prefer to publish your web to a local directory (possibly even in place with your Org files) and then use unison or rsync to do the synchronization with the remote host.

Since Unison (for example) can be configured as to which files to transfer to a certain remote destination, it can greatly simplify the project publishing definition. Simply keep all files in the correct location, process your Org files with org-publish and let the synchronization tool do the rest. You do not need, in this scenario, to include attachments such as jpg, css or gif files in the project definition since the 3rd party tool syncs them.

Publishing to a local directory is also much faster than to a remote one, so that you can afford more easily to republish entire projects. If you set org-publish-use-timestamps-flag to nil, you gain the main benefit of re-including any changed external files such as source example files you might include with #+INCLUDE:. The timestamp mechanism in Org is not smart enough to detect if included files have been modified.

Next: , Previous: Uploading files, Up: Publishing

13.3 Sample configuration

Below we provide two example configurations. The first one is a simple project publishing only a set of Org files. The second example is more complex, with a multi-component project.

Next: , Previous: Sample configuration, Up: Sample configuration

13.3.1 Example: simple publishing configuration

This example publishes a set of Org files to the public_html directory on the local machine.

     (setq org-publish-project-alist
              :base-directory "~/org/"
              :publishing-directory "~/public_html"
              :section-numbers nil
              :with-toc nil
              :html-head "<link rel=\"stylesheet\"

Previous: Simple example, Up: Sample configuration

13.3.2 Example: complex publishing configuration

This more complicated example publishes an entire website, including Org files converted to HTML, image files, Emacs Lisp source code, and style sheets. The publishing directory is remote and private files are excluded.

To ensure that links are preserved, care should be taken to replicate your directory structure on the web server, and to use relative file paths. For example, if your Org files are kept in ~/org and your publishable images in ~/images, you would link to an image with


On the web server, the relative path to the image should be the same. You can accomplish this by setting up an "images" folder in the right place on the web server, and publishing images to it.

     (setq org-publish-project-alist
               :base-directory "~/org/"
               :base-extension "org"
               :publishing-directory "/ssh:user@host:~/html/notebook/"
               :publishing-function org-html-publish-to-html
               :exclude ""   ;; regexp
               :headline-levels 3
               :section-numbers nil
               :with-toc nil
               :html-head "<link rel=\"stylesheet\"
                       href=\"../other/mystyle.css\" type=\"text/css\"/>"
               :html-preamble t)
               :base-directory "~/images/"
               :base-extension "jpg\\|gif\\|png"
               :publishing-directory "/ssh:user@host:~/html/images/"
               :publishing-function org-publish-attachment)
               :base-directory "~/other/"
               :base-extension "css\\|el"
               :publishing-directory "/ssh:user@host:~/html/other/"
               :publishing-function org-publish-attachment)
              ("website" :components ("orgfiles" "images" "other"))))

Previous: Sample configuration, Up: Publishing

13.4 Triggering publication

Once properly configured, Org can publish with the following commands:

C-c C-e P x     (org-publish)
Prompt for a specific project and publish all files that belong to it.
C-c C-e P p     (org-publish-current-project)
Publish the project containing the current file.
C-c C-e P f     (org-publish-current-file)
Publish only the current file.
C-c C-e P a     (org-publish-all)
Publish every project.

Org uses timestamps to track when a file has changed. The above functions normally only publish changed files. You can override this and force publishing of all files by giving a prefix argument to any of the commands above, or by customizing the variable org-publish-use-timestamps-flag. This may be necessary in particular if files include other files via #+SETUPFILE: or #+INCLUDE:.

Next: , Previous: Publishing, Up: Top

14 Working with source code

Source code can be included in Org mode documents using a ‘src’ block, e.g.:

     #+BEGIN_SRC emacs-lisp
       (defun org-xor (a b)
          "Exclusive or."
          (if a (not b) b))

Org mode provides a number of features for working with live source code, including editing of code blocks in their native major-mode, evaluation of code blocks, converting code blocks into source files (known as tangling in literate programming), and exporting code blocks and their results in several formats. This functionality was contributed by Eric Schulte and Dan Davison, and was originally named Org-babel.

The following sections describe Org mode's code block handling facilities.

Next: , Previous: Working With Source Code, Up: Working With Source Code

14.1 Structure of code blocks

Live code blocks can be specified with a ‘src’ block or inline.168 The structure of a ‘src’ block is

     #+NAME: <name>
     #+BEGIN_SRC <language> <switches> <header arguments>

The #+NAME: line is optional, and can be used to name the code block. Live code blocks require that a language be specified on the #+BEGIN_SRC line. Switches and header arguments are optional. Live code blocks can also be specified inline using



     src_<language>[<header arguments>]{<body>}
<#+NAME: name>
This line associates a name with the code block. This is similar to the #+NAME: Name lines that can be used to name tables in Org mode files. Referencing the name of a code block makes it possible to evaluate the block from other places in the file, from other files, or from Org mode table formulas (see The spreadsheet). Names are assumed to be unique and the behavior of Org mode when two or more blocks share the same name is undefined.
The language of the code in the block (see Languages).
Optional switches control code block export (see the discussion of switches in Literal examples)
<header arguments>
Optional header arguments control many aspects of evaluation, export and tangling of code blocks (see Header arguments). Header arguments can also be set on a per-buffer or per-subtree basis using properties.
source code, header arguments
Source code in the specified language.

Next: , Previous: Structure of code blocks, Up: Working With Source Code

14.2 Editing source code

Use C-c ' to edit the current code block. This brings up a language major-mode edit buffer containing the body of the code block. Manually saving this buffer with <C-x C-s> will write the contents back to the Org buffer. You can also set org-edit-src-auto-save-idle-delay to save the base buffer after some idle delay, or org-edit-src-turn-on-auto-save to auto-save this buffer into a separate file using auto-save-mode. Use C-c ' again to exit.

The org-src-mode minor mode will be active in the edit buffer. The following variables can be used to configure the behavior of the edit buffer. See also the customization group org-edit-structure for further configuration options.

If an Emacs major-mode named <lang>-mode exists, where <lang> is the language named in the header line of the code block, then the edit buffer will be placed in that major-mode. This variable can be used to map arbitrary language names to existing major modes.
Controls the way Emacs windows are rearranged when the edit buffer is created.
By default, the value is nil, which means that when code blocks are evaluated during export or tangled, they are re-inserted into the code block, which may replace sequences of spaces with tab characters. When non-nil, whitespace in code blocks will be preserved during export or tangling, exactly as it appears. This variable is especially useful for tangling languages such as Python, in which whitespace indentation in the output is critical.
By default, Org will ask before returning to an open edit buffer. Set this variable to nil to switch without asking.

To turn on native code fontification in the Org buffer, configure the variable org-src-fontify-natively.

Next: , Previous: Editing source code, Up: Working With Source Code

14.3 Exporting code blocks

It is possible to export the code of code blocks, the results of code block evaluation, both the code and the results of code block evaluation, or none. For most languages, the default exports code. However, for some languages (e.g., ditaa) the default exports the results of code block evaluation. For information on exporting code block bodies, see Literal examples.

The :exports header argument can be used to specify export behavior:

Header arguments:
:exports code
The default in most languages. The body of the code block is exported, as described in Literal examples.
:exports results
The code block will be evaluated and the results will be placed in the Org mode buffer for export, either updating previous results of the code block located anywhere in the buffer or, if no previous results exist, placing the results immediately after the code block. The body of the code block will not be exported.
:exports both
Both the code block and its results will be exported.
:exports none
Neither the code block nor its results will be exported.

It is possible to inhibit the evaluation of code blocks during export. Setting the org-export-babel-evaluate variable to nil will ensure that no code blocks are evaluated as part of the export process. This can be useful in situations where potentially untrusted Org mode files are exported in an automated fashion, for example when Org mode is used as the markup language for a wiki. It is also possible to set this variable to 'inline-only. In that case, only inline code blocks will be evaluated, in order to insert their results. Non-inline code blocks are assumed to have their results already inserted in the buffer by manual evaluation. This setting is useful to avoid expensive recalculations during export, not to provide security.

Next: , Previous: Exporting code blocks, Up: Working With Source Code

14.4 Extracting source code

Creating pure source code files by extracting code from source blocks is referred to as “tangling”—a term adopted from the literate programming community. During “tangling” of code blocks their bodies are expanded using org-babel-expand-src-block which can expand both variable and “noweb” style references (see Noweb reference syntax).

Header arguments
:tangle no
The default. The code block is not included in the tangled output.
:tangle yes
Include the code block in the tangled output. The output file name is the name of the org file with the extension ‘.org’ replaced by the extension for the block language.
:tangle filename
Include the code block in the tangled output to file ‘filename’.

Tangle the current file. Bound to C-c C-v t.

With prefix argument only tangle the current code block.

Choose a file to tangle. Bound to C-c C-v f.
This hook is run from within code files tangled by org-babel-tangle. Example applications could include post-processing, compilation or evaluation of tangled code files.
Jumping between code and Org

When tangling code from an Org-mode buffer to a source code file, you'll frequently find yourself viewing the file of tangled source code (e.g., many debuggers point to lines of the source code file). It is useful to be able to navigate from the tangled source to the Org-mode buffer from which the code originated.

The org-babel-tangle-jump-to-org function provides this jumping from code to Org-mode functionality. Two header arguments are required for jumping to work, first the padline (padline) option must be set to true (the default setting), second the comments (comments) header argument must be set to links, which will insert comments into the source code buffer which point back to the original Org-mode file.

Next: , Previous: Extracting source code, Up: Working With Source Code

14.5 Evaluating code blocks

Code blocks can be evaluated169 and the results of evaluation optionally placed in the Org mode buffer. The results of evaluation are placed following a line that begins by default with #+RESULTS and optionally a cache identifier and/or the name of the evaluated code block. The default value of #+RESULTS can be changed with the customizable variable org-babel-results-keyword.

By default, the evaluation facility is only enabled for Lisp code blocks specified as emacs-lisp. However, source code blocks in many languages can be evaluated within Org mode (see Languages for a list of supported languages and Structure of code blocks for information on the syntax used to define a code block).

There are a number of ways to evaluate code blocks. The simplest is to press C-c C-c or C-c C-v e with the point on a code block170. This will call the org-babel-execute-src-block function to evaluate the block and insert its results into the Org mode buffer. It is also possible to evaluate named code blocks from anywhere in an Org mode buffer or an Org mode table. Live code blocks located in the current Org mode buffer or in the “Library of Babel” (see Library of Babel) can be executed. Named code blocks can be executed with a separate #+CALL: line or inline within a block of text.

The syntax of the #+CALL: line is

     #+CALL: <name>(<arguments>)
     #+CALL: <name>[<inside header arguments>](<arguments>) <end header arguments>

The syntax for inline evaluation of named code blocks is

     ... call_<name>(<arguments>) ...
     ... call_<name>[<inside header arguments>](<arguments>)[<end header arguments>] ...
The name of the code block to be evaluated (see Structure of code blocks).
Arguments specified in this section will be passed to the code block. These arguments use standard function call syntax, rather than header argument syntax. For example, a #+CALL: line that passes the number four to a code block named double, which declares the header argument :var n=2, would be written as #+CALL: double(n=4).
<inside header arguments>
Inside header arguments are passed through and applied to the named code block. These arguments use header argument syntax rather than standard function call syntax. Inside header arguments affect how the code block is evaluated. For example, [:results output] will collect the results of everything printed to STDOUT during execution of the code block.
<end header arguments>
End header arguments are applied to the calling instance and do not affect evaluation of the named code block. They affect how the results are incorporated into the Org mode buffer and how the call line is exported. For example, :results html will insert the results of the call line evaluation in the Org buffer, wrapped in a BEGIN_HTML: block.

For more examples of passing header arguments to #+CALL: lines see Header arguments in function calls.

Next: , Previous: Evaluating code blocks, Up: Working With Source Code

14.6 Library of Babel

The “Library of Babel” consists of code blocks that can be called from any Org mode file. Code blocks defined in the “Library of Babel” can be called remotely as if they were in the current Org mode buffer (see Evaluating code blocks for information on the syntax of remote code block evaluation).

The central repository of code blocks in the “Library of Babel” is housed in an Org mode file located in the ‘contrib’ directory of Org mode.

Users can add code blocks they believe to be generally useful to their “Library of Babel.” The code blocks can be stored in any Org mode file and then loaded into the library with org-babel-lob-ingest.

Code blocks located in any Org mode file can be loaded into the “Library of Babel” with the org-babel-lob-ingest function, bound to C-c C-v i.

Next: , Previous: Library of Babel, Up: Working With Source Code

14.7 Languages

Code blocks in the following languages are supported.

Language Identifier Language Identifier
Asymptote asymptote Awk awk
Emacs Calc calc C C
C++ C++ Clojure clojure
CSS css ditaa ditaa
Graphviz dot Emacs Lisp emacs-lisp
gnuplot gnuplot Haskell haskell
Java java
Javascript js LaTeX latex
Ledger ledger Lisp lisp
Lilypond lilypond MATLAB matlab
Mscgen mscgen Objective Caml ocaml
Octave octave Org mode org
Oz oz Perl perl
Plantuml plantuml Python python
R R Ruby ruby
Sass sass Scheme scheme
GNU Screen screen shell sh
SQL sql SQLite sqlite

Language-specific documentation is available for some languages. If available, it can be found at

The option org-babel-load-languages controls which languages are enabled for evaluation (by default only emacs-lisp is enabled). This variable can be set using the customization interface or by adding code like the following to your emacs configuration.

The following disables emacs-lisp evaluation and enables evaluation of R code blocks.
      '((emacs-lisp . nil)
        (R . t)))

It is also possible to enable support for a language by loading the related elisp file with require.

The following adds support for evaluating clojure code blocks.
     (require 'ob-clojure)

Next: , Previous: Languages, Up: Working With Source Code

14.8 Header arguments

Code block functionality can be configured with header arguments. This section provides an overview of the use of header arguments, and then describes each header argument in detail.

Next: , Previous: Header arguments, Up: Header arguments

14.8.1 Using header arguments

The values of header arguments can be set in several way. When the header arguments in each layer have been determined, they are combined in order from the first, least specific (having the lowest priority) up to the last, most specific (having the highest priority). A header argument with a higher priority replaces the same header argument specified at lower priority.

Next: , Previous: Using header arguments, Up: Using header arguments

System-wide header arguments

System-wide values of header arguments can be specified by adapting the org-babel-default-header-args variable:

     :session    => "none"
     :results    => "replace"
     :exports    => "code"
     :cache      => "no"
     :noweb      => "no"

For example, the following example could be used to set the default value of :noweb header arguments to yes. This would have the effect of expanding :noweb references by default when evaluating source code blocks.

     (setq org-babel-default-header-args
           (cons '(:noweb . "yes")
                 (assq-delete-all :noweb org-babel-default-header-args)))

Next: , Previous: System-wide header arguments, Up: Using header arguments

Language-specific header arguments

Each language can define its own set of default header arguments in variable org-babel-default-header-args:<lang>, where <lang> is the name of the language. See the language-specific documentation available online at

Next: , Previous: Language-specific header arguments, Up: Using header arguments

Header arguments in Org mode properties

Buffer-wide header arguments may be specified as properties through the use of #+PROPERTY: lines placed anywhere in an Org mode file (see Property syntax).

For example the following would set session to *R* (only for R code blocks), and results to silent for every code block in the buffer, ensuring that all execution took place in the same session, and no results would be inserted into the buffer.

     #+PROPERTY: header-args:R  :session *R*
     #+PROPERTY: header-args    :results silent

Header arguments read from Org mode properties can also be set on a per-subtree basis using property drawers (see Property syntax). When properties are used to set default header arguments, they are always looked up with inheritance, regardless of the value of org-use-property-inheritance. Properties are evaluated as seen by the outermost call or source block.171

In the following example the value of the :cache header argument will default to yes in all code blocks in the subtree rooted at the following heading:

     * outline header
       :header-args:    :cache yes

Properties defined in this way override the properties set in org-babel-default-header-args and are applied for all activated languages. It is convenient to use the org-set-property function bound to C-c C-x p to set properties in Org mode documents.

Next: , Previous: Header arguments in Org mode properties, Up: Using header arguments

Language-specific header arguments in Org mode properties

Language-specific header arguments are also read from properties header-args:<lang> where <lang> is the name of the language targeted. As an example

     * Heading
       :header-args:clojure:    :session *clojure-1*
       :header-args:R:          :session *R*
     ** Subheading
       :header-args:clojure:    :session *clojure-2*

would independently set a default session header argument for R and clojure for calls and source blocks under subtree “Heading” and change to a different clojure setting for evaluations under subtree “Subheading”, while the R session is inherited from “Heading” and therefore unchanged.

Next: , Previous: Language-specific header arguments in Org mode properties, Up: Using header arguments

Code block specific header arguments

The most common way to assign values to header arguments is at the code block level. This can be done by listing a sequence of header arguments and their values as part of the #+BEGIN_SRC line. Properties set in this way override both the values of org-babel-default-header-args and header arguments specified as properties. In the following example, the :results header argument is set to silent, meaning the results of execution will not be inserted in the buffer, and the :exports header argument is set to code, meaning only the body of the code block will be preserved on export to HTML or LaTeX.

     #+NAME: factorial
     #+BEGIN_SRC haskell :results silent :exports code :var n=0
     fac 0 = 1
     fac n = n * fac (n-1)

Similarly, it is possible to set header arguments for inline code blocks

     src_haskell[:exports both]{fac 5}

Code block header arguments can span multiple lines using #+HEADER: or #+HEADERS: lines preceding a code block or nested between the #+NAME: line and the #+BEGIN_SRC line of a named code block. Multi-line header arguments on an un-named code block:

      #+HEADERS: :var data1=1
      #+BEGIN_SRC emacs-lisp :var data2=2
        (message "data1:%S, data2:%S" data1 data2)
      : data1:1, data2:2

Multi-line header arguments on a named code block:

        #+NAME: named-block
        #+HEADER: :var data=2
        #+BEGIN_SRC emacs-lisp
          (message "data:%S" data)
        #+RESULTS: named-block
        : data:2

Previous: Code block specific header arguments, Up: Using header arguments

Header arguments in function calls

At the most specific level, header arguments for “Library of Babel” or #+CALL: lines can be set as shown in the two examples below. For more information on the structure of #+CALL: lines see Evaluating code blocks.

The following will apply the :exports results header argument to the evaluation of the #+CALL: line.

     #+CALL: factorial(n=5) :exports results

The following will apply the :session special header argument to the evaluation of the factorial code block.

     #+CALL: factorial[:session special](n=5)

Previous: Using header arguments, Up: Header arguments

14.8.2 Specific header arguments

Header arguments consist of an initial colon followed by the name of the argument in lowercase letters. The following header arguments are defined:

Additional header arguments are defined on a language-specific basis, see Languages.

Next: , Previous: Specific header arguments, Up: Specific header arguments :var

The :var header argument is used to pass arguments to code blocks. The specifics of how arguments are included in a code block vary by language; these are addressed in the language-specific documentation. However, the syntax used to specify arguments is the same across all languages. In every case, variables require a default value when they are declared.

The values passed to arguments can either be literal values, references, or Emacs Lisp code (see Emacs Lisp evaluation of variables). References include anything in the Org mode file that takes a #+NAME: or #+RESULTS: line: tables, lists, #+BEGIN_EXAMPLE blocks, other code blocks and the results of other code blocks.

Note: When a reference is made to another code block, the referenced block will be evaluated unless it has current cached results (see cache).

Argument values can be indexed in a manner similar to arrays (see Indexable variable values).

The following syntax is used to pass arguments to code blocks using the :var header argument.

     :var name=assign

The argument, assign, can either be a literal value, such as a string ‘"string"’ or a number ‘9’, or a reference to a table, a list, a literal example, another code block (with or without arguments), or the results of evaluating another code block.

Here are examples of passing values by reference:

an Org mode table named with either a #+NAME: line
          #+NAME: example-table
          | 1 |
          | 2 |
          | 3 |
          | 4 |
          #+NAME: table-length
          #+BEGIN_SRC emacs-lisp :var table=example-table
          (length table)
          #+RESULTS: table-length
          : 4

a simple list named with a #+NAME: line (note that nesting is not carried through to the source code block)
          #+NAME: example-list
            - simple
              - not
              - nested
            - list
          #+BEGIN_SRC emacs-lisp :var x=example-list
            (print x)
          | simple | list |

code block without arguments
a code block name (from the example above), as assigned by #+NAME:, optionally followed by parentheses
          #+BEGIN_SRC emacs-lisp :var length=table-length()
          (* 2 length)
          : 8

code block with arguments
a code block name, as assigned by #+NAME:, followed by parentheses and optional arguments passed within the parentheses following the code block name using standard function call syntax
          #+NAME: double
          #+BEGIN_SRC emacs-lisp :var input=8
          (* 2 input)
          #+RESULTS: double
          : 16
          #+NAME: squared
          #+BEGIN_SRC emacs-lisp :var input=double(input=1)
          (* input input)
          #+RESULTS: squared
          : 4

literal example
a literal example block named with a #+NAME: line
          #+NAME: literal-example
          A literal example
          on two lines
          #+NAME: read-literal-example
          #+BEGIN_SRC emacs-lisp :var x=literal-example
            (concatenate 'string x " for you.")
          #+RESULTS: read-literal-example
          : A literal example
          : on two lines for you.
Indexable variable values

It is possible to reference portions of variable values by “indexing” into the variables. Indexes are 0 based with negative values counting back from the end. If an index is separated by ,s then each subsequent section will index into the next deepest nesting or dimension of the value. Note that this indexing occurs before other table related header arguments like :hlines, :colnames and :rownames are applied. The following example assigns the last cell of the first row the table example-table to the variable data:

     #+NAME: example-table
     | 1 | a |
     | 2 | b |
     | 3 | c |
     | 4 | d |
     #+BEGIN_SRC emacs-lisp :var data=example-table[0,-1]
     : a

Ranges of variable values can be referenced using two integers separated by a :, in which case the entire inclusive range is referenced. For example the following assigns the middle three rows of example-table to data.

     #+NAME: example-table
     | 1 | a |
     | 2 | b |
     | 3 | c |
     | 4 | d |
     | 5 | 3 |
     #+BEGIN_SRC emacs-lisp :var data=example-table[1:3]
     | 2 | b |
     | 3 | c |
     | 4 | d |

Additionally, an empty index, or the single character *, are both interpreted to mean the entire range and as such are equivalent to 0:-1, as shown in the following example in which the entire first column is referenced.

     #+NAME: example-table
     | 1 | a |
     | 2 | b |
     | 3 | c |
     | 4 | d |
     #+BEGIN_SRC emacs-lisp :var data=example-table[,0]
     | 1 | 2 | 3 | 4 |

It is possible to index into the results of code blocks as well as tables. Any number of dimensions can be indexed. Dimensions are separated from one another by commas, as shown in the following example.

     #+NAME: 3D
     #+BEGIN_SRC emacs-lisp
       '(((1  2  3)  (4  5  6)  (7  8  9))
         ((10 11 12) (13 14 15) (16 17 18))
         ((19 20 21) (22 23 24) (25 26 27)))
     #+BEGIN_SRC emacs-lisp :var data=3D[1,,1]
     | 11 | 14 | 17 |
Emacs Lisp evaluation of variables

Emacs lisp code can be used to initialize variable values. When a variable value starts with (, [, ' or ` it will be evaluated as Emacs Lisp and the result of the evaluation will be assigned as the variable value. The following example demonstrates use of this evaluation to reliably pass the file-name of the Org mode buffer to a code block—note that evaluation of header arguments is guaranteed to take place in the original Org mode file, while there is no such guarantee for evaluation of the code block body.

     #+BEGIN_SRC sh :var filename=(buffer-file-name) :exports both
       wc -w $filename

Note that values read from tables and lists will not be evaluated as Emacs Lisp, as shown in the following example.

     #+NAME: table
     | (a b c) |
     #+HEADERS: :var data=table[0,0]
     #+BEGIN_SRC perl
     : (a b c)

Next: , Previous: var, Up: Specific header arguments :results

There are four classes of :results header argument. Only one option per class may be supplied per code block.


The following options are mutually exclusive, and specify how the results should be collected from the code block.


The following options are mutually exclusive and specify what type of results the code block will return. By default, results are inserted as either a table or scalar depending on their value.


The following options are mutually exclusive and specify what type of results the code block will return. By default, results are inserted according to the type as specified above.


The following results options indicate what happens with the results once they are collected.

Next: , Previous: results, Up: Specific header arguments :file

The header argument :file is used to specify an external file in which to save code block results. After code block evaluation an Org mode style [[file:]] link (see Link format) to the file will be inserted into the Org mode buffer. Some languages including R, gnuplot, dot, and ditaa provide special handling of the :file header argument automatically wrapping the code block body in the boilerplate code required to save output to the specified file. This is often useful for saving graphical output of a code block to the specified file.

The argument to :file should be either a string specifying the path to a file, or a list of two strings in which case the first element of the list should be the path to a file and the second a description for the link.

Next: , Previous: file, Up: Specific header arguments :file-desc

The value of the :file-desc header argument is used to provide a description for file code block results which are inserted as Org mode links (see Link format). If the :file-desc header argument is given with no value the link path will be placed in both the “link” and the “description” portion of the Org mode link.

Next: , Previous: file-desc, Up: Specific header arguments :dir and remote execution

While the :file header argument can be used to specify the path to the output file, :dir specifies the default directory during code block execution. If it is absent, then the directory associated with the current buffer is used. In other words, supplying :dir path temporarily has the same effect as changing the current directory with M-x cd path RET, and then not supplying :dir. Under the surface, :dir simply sets the value of the Emacs variable default-directory.

When using :dir, you should supply a relative path for file output (e.g., :file myfile.jpg or :file results/myfile.jpg) in which case that path will be interpreted relative to the default directory.

In other words, if you want your plot to go into a folder called Work in your home directory, you could use

     #+BEGIN_SRC R :file myplot.png :dir ~/Work
     matplot(matrix(rnorm(100), 10), type="l")
Remote execution

A directory on a remote machine can be specified using tramp file syntax, in which case the code will be evaluated on the remote machine. An example is

     #+BEGIN_SRC R :file plot.png :dir /
     plot(1:10, main=system("hostname", intern=TRUE))

Text results will be returned to the local Org mode buffer as usual, and file output will be created on the remote machine with relative paths interpreted relative to the remote directory. An Org mode link to the remote file will be created.

So, in the above example a plot will be created on the remote machine, and a link of the following form will be inserted in the org buffer:


Most of this functionality follows immediately from the fact that :dir sets the value of the Emacs variable default-directory, thanks to tramp. Those using XEmacs, or GNU Emacs prior to version 23 may need to install tramp separately in order for these features to work correctly.

Further points

Next: , Previous: dir, Up: Specific header arguments :exports

The :exports header argument specifies what should be included in HTML or LaTeX exports of the Org mode file.

Next: , Previous: exports, Up: Specific header arguments :tangle

The :tangle header argument specifies whether or not the code block should be included in tangled extraction of source code files.

Next: , Previous: tangle, Up: Specific header arguments :mkdirp

The :mkdirp header argument can be used to create parent directories of tangled files when missing. This can be set to yes to enable directory creation or to no to inhibit directory creation.

Next: , Previous: mkdirp, Up: Specific header arguments :comments

By default code blocks are tangled to source-code files without any insertion of comments beyond those which may already exist in the body of the code block. The :comments header argument can be set as follows to control the insertion of extra comments into the tangled code file.

Next: , Previous: comments, Up: Specific header arguments :padline

Control in insertion of padding lines around code block bodies in tangled code files. The default value is yes which results in insertion of newlines before and after each tangled code block. The following arguments are accepted.

Next: , Previous: padline, Up: Specific header arguments :no-expand

By default, code blocks are expanded with org-babel-expand-src-block during tangling. This has the effect of assigning values to variables specified with :var (see var), and of replacing “noweb” references (see Noweb reference syntax) with their targets. The :no-expand header argument can be used to turn off this behavior.

Next: , Previous: no-expand, Up: Specific header arguments :session

The :session header argument starts a session for an interpreted language where state is preserved.

By default, a session is not started.

A string passed to the :session header argument will give the session a name. This makes it possible to run concurrent sessions for each interpreted language.

Next: , Previous: session, Up: Specific header arguments :noweb

The :noweb header argument controls expansion of “noweb” syntax references (see Noweb reference syntax) when the code block is evaluated, tangled, or exported. The :noweb header argument can have one of the five values: no, yes, tangle, or no-export strip-export.

Noweb prefix lines

Noweb insertions are now placed behind the line prefix of the <<reference>>. This behavior is illustrated in the following example. Because the <<example>> noweb reference appears behind the SQL comment syntax, each line of the expanded noweb reference will be commented.

This code block:

     -- <<example>>

expands to:

     -- this is the
     -- multi-line body of example

Note that noweb replacement text that does not contain any newlines will not be affected by this change, so it is still possible to use inline noweb references.

Next: , Previous: noweb, Up: Specific header arguments :noweb-ref

When expanding “noweb” style references the bodies of all code block with either a block name matching the reference name or a :noweb-ref header argument matching the reference name will be concatenated together to form the replacement text.

By setting this header argument at the sub-tree or file level, simple code block concatenation may be achieved. For example, when tangling the following Org mode file, the bodies of code blocks will be concatenated into the resulting pure code file172.

      #+BEGIN_SRC sh :tangle yes :noweb yes :shebang #!/bin/sh
      * the mount point of the fullest disk
        :noweb-ref: fullest-disk
      ** query all mounted disks
      #+BEGIN_SRC sh
        df \
      ** strip the header row
      #+BEGIN_SRC sh
        |sed '1d' \
      ** sort by the percent full
      #+BEGIN_SRC sh
        |awk '{print $5 " " $6}'|sort -n |tail -1 \
      ** extract the mount point
      #+BEGIN_SRC sh
        |awk '{print $2}'

The :noweb-sep (see noweb-sep) header argument holds the string used to separate accumulate noweb references like those above. By default a newline is used.

Next: , Previous: noweb-ref, Up: Specific header arguments :noweb-sep

The :noweb-sep header argument holds the string used to separate accumulate noweb references (see noweb-ref). By default a newline is used.

Next: , Previous: noweb-sep, Up: Specific header arguments :cache

The :cache header argument controls the use of in-buffer caching of the results of evaluating code blocks. It can be used to avoid re-evaluating unchanged code blocks. Note that the :cache header argument will not attempt to cache results when the :session header argument is used, because the results of the code block execution may be stored in the session outside of the Org mode buffer. The :cache header argument can have one of two values: yes or no.

Code block caches notice if the value of a variable argument to the code block has changed. If this is the case, the cache is invalidated and the code block is re-run. In the following example, caller will not be re-run unless the results of random have changed since it was last run.

      #+NAME: random
      #+BEGIN_SRC R :cache yes
      #+RESULTS[a2a72cd647ad44515fab62e144796432793d68e1]: random
      #+NAME: caller
      #+BEGIN_SRC emacs-lisp :var x=random :cache yes
      #+RESULTS[bec9c8724e397d5df3b696502df3ed7892fc4f5f]: caller

Next: , Previous: cache, Up: Specific header arguments :sep

The :sep header argument can be used to control the delimiter used when writing tabular results out to files external to Org mode. This is used either when opening tabular results of a code block by calling the org-open-at-point function bound to C-c C-o on the code block, or when writing code block results to an external file (see file) header argument.

By default, when :sep is not specified output tables are tab delimited.

Next: , Previous: sep, Up: Specific header arguments :hlines

Tables are frequently represented with one or more horizontal lines, or hlines. The :hlines argument to a code block accepts the values yes or no, with a default value of no.

Next: , Previous: hlines, Up: Specific header arguments :colnames

The :colnames header argument accepts the values yes, no, or nil for unassigned. The default value is nil. Note that the behavior of the :colnames header argument may differ across languages.

Next: , Previous: colnames, Up: Specific header arguments :rownames

The :rownames header argument can take on the values yes or no, with a default value of no. Note that Emacs Lisp code blocks ignore the :rownames header argument entirely given the ease with which tables with row names may be handled directly in Emacs Lisp.

Next: , Previous: rownames, Up: Specific header arguments :shebang

Setting the :shebang header argument to a string value (e.g., :shebang "#!/bin/bash") causes the string to be inserted as the first line of any tangled file holding the code block, and the file permissions of the tangled file are set to make it executable.

Next: , Previous: shebang, Up: Specific header arguments :tangle-mode

The tangle-mode header argument controls the permission set on tangled files. The value of this header argument will be passed to set-file-modes. For example, to set a tangled file as read only use :tangle-mode (identity #o444), or to set a tangled file as executable use :tangle-mode (identity #o755). Blocks with shebang (shebang) header arguments will automatically be made executable unless the tangle-mode header argument is also used. The behavior is undefined if multiple code blocks with different values for the tangle-mode header argument are tangled to the same file.

Next: , Previous: tangle-mode, Up: Specific header arguments :eval

The :eval header argument can be used to limit the evaluation of specific code blocks. The :eval header argument can be useful for protecting against the evaluation of dangerous code blocks or to ensure that evaluation will require a query regardless of the value of the org-confirm-babel-evaluate variable. The possible values of :eval and their effects are shown below.

never or no
The code block will not be evaluated under any circumstances.
Evaluation of the code block will require a query.
never-export or no-export
The code block will not be evaluated during export but may still be called interactively.
Evaluation of the code block during export will require a query.

If this header argument is not set then evaluation is determined by the value of the org-confirm-babel-evaluate variable see Code evaluation security.

Next: , Previous: eval, Up: Specific header arguments :wrap

The :wrap header argument is used to mark the results of source block evaluation. The header argument can be passed a string that will be appended to #+BEGIN_ and #+END_, which will then be used to wrap the results. If not string is specified then the results will be wrapped in a #+BEGIN/END_RESULTS block.

Next: , Previous: wrap, Up: Specific header arguments :post

The :post header argument is used to post-process the results of a code block execution. When a post argument is given, the results of the code block will temporarily be bound to the *this* variable. This variable may then be included in header argument forms such as those used in var header argument specifications allowing passing of results to other code blocks, or direct execution via Emacs Lisp.

The following example illustrates the usage of the :post header argument.

     #+name: attr_wrap
     #+begin_src sh :var data="" :var width="\\textwidth" :results output
       echo "#+ATTR_LATEX :width $width"
       echo "$data"
     #+header: :file /tmp/it.png
     #+begin_src dot :post attr_wrap(width="5cm", data=*this*) :results drawer
               a -> b;
               b -> c;
               c -> a;
     #+ATTR_LATEX :width 5cm

Next: , Previous: post, Up: Specific header arguments :prologue

The value of the prologue header argument will be prepended to the code block body before execution. For example, :prologue "reset" may be used to reset a gnuplot session before execution of a particular code block, or the following configuration may be used to do this for all gnuplot code blocks. Also see epilogue.

     (add-to-list 'org-babel-default-header-args:gnuplot
                  '((:prologue . "reset")))

Previous: prologue, Up: Specific header arguments :epilogue

The value of the epilogue header argument will be appended to the code block body before execution. Also see prologue.

Next: , Previous: Header arguments, Up: Working With Source Code

14.9 Results of evaluation

The way in which results are handled depends on whether a session is invoked, as well as on whether :results value or :results output is used. The following table shows the table possibilities. For a full listing of the possible results header arguments see results.

Non-session Session
:results value value of last expression value of last expression
:results output contents of STDOUT concatenation of interpreter output

Note: With :results value, the result in both :session and non-session is returned to Org mode as a table (a one- or two-dimensional vector of strings or numbers) when appropriate.

14.9.1 Non-session :results value

This is the default. Internally, the value is obtained by wrapping the code in a function definition in the external language, and evaluating that function. Therefore, code should be written as if it were the body of such a function. In particular, note that Python does not automatically return a value from a function unless a return statement is present, and so a ‘return’ statement will usually be required in Python.

This is the only one of the four evaluation contexts in which the code is automatically wrapped in a function definition. :results output

The code is passed to the interpreter as an external process, and the contents of the standard output stream are returned as text. (In certain languages this also contains the error output stream; this is an area for future work.)

14.9.2 Session :results value

The code is passed to an interpreter running as an interactive Emacs inferior process. Only languages which provide tools for interactive evaluation of code have session support, so some language (e.g., C and ditaa) do not support the :session header argument, and in other languages (e.g., Python and Haskell) which have limitations on the code which may be entered into interactive sessions, those limitations apply to the code in code blocks using the :session header argument as well.

Unless the :results output option is supplied (see below) the result returned is the result of the last evaluation performed by the interpreter. (This is obtained in a language-specific manner: the value of the variable _ in Python and Ruby, and the value of .Last.value in R). :results output

The code is passed to the interpreter running as an interactive Emacs inferior process. The result returned is the concatenation of the sequence of (text) output from the interactive interpreter. Notice that this is not necessarily the same as what would be sent to STDOUT if the same code were passed to a non-interactive interpreter running as an external process. For example, compare the following two blocks:

     #+BEGIN_SRC python :results output
      print "hello"
      print "bye"
     : hello
     : bye

In non-session mode, the “2” is not printed and does not appear.

     #+BEGIN_SRC python :results output :session
      print "hello"
      print "bye"
     : hello
     : 2
     : bye

But in :session mode, the interactive interpreter receives input “2” and prints out its value, “2”. (Indeed, the other print statements are unnecessary here).

Next: , Previous: Results of evaluation, Up: Working With Source Code

14.10 Noweb reference syntax

The “noweb” (see Literate Programming system allows named blocks of code to be referenced by using the familiar Noweb syntax:


When a code block is tangled or evaluated, whether or not “noweb” references are expanded depends upon the value of the :noweb header argument. If :noweb yes, then a Noweb reference is expanded before evaluation. If :noweb no, the default, then the reference is not expanded before evaluation. See the noweb-ref header argument for a more flexible way to resolve noweb references.

It is possible to include the results of a code block rather than the body. This is done by appending parenthesis to the code block name which may optionally contain arguments to the code block as shown below.

     <<code-block-name(optional arguments)>>

Note: the default value, :noweb no, was chosen to ensure that correct code is not broken in a language, such as Ruby, where <<arg>> is a syntactically valid construct. If <<arg>> is not syntactically valid in languages that you use, then please consider setting the default value.

Note: if noweb tangling is slow in large Org mode files consider setting the org-babel-use-quick-and-dirty-noweb-expansion variable to t. This will result in faster noweb reference resolution at the expense of not correctly resolving inherited values of the :noweb-ref header argument.

Next: , Previous: Noweb reference syntax, Up: Working With Source Code

14.11 Key bindings and useful functions

Many common Org mode key sequences are re-bound depending on the context.

Within a code block, the following key bindings are active:

C-c C-c org-babel-execute-src-block
C-c C-o org-babel-open-src-block-result
M-<up> org-babel-load-in-session
M-<down> org-babel-switch-to-session

In an Org mode buffer, the following key bindings are active:

C-c C-v p or C-c C-v C-p org-babel-previous-src-block
C-c C-v n or C-c C-v C-n org-babel-next-src-block
C-c C-v e or C-c C-v C-e org-babel-execute-maybe
C-c C-v o or C-c C-v C-o org-babel-open-src-block-result
C-c C-v v or C-c C-v C-v org-babel-expand-src-block
C-c C-v u or C-c C-v C-u org-babel-goto-src-block-head
C-c C-v g or C-c C-v C-g org-babel-goto-named-src-block
C-c C-v r or C-c C-v C-r org-babel-goto-named-result
C-c C-v b or C-c C-v C-b org-babel-execute-buffer
C-c C-v s or C-c C-v C-s org-babel-execute-subtree
C-c C-v d or C-c C-v C-d org-babel-demarcate-block
C-c C-v t or C-c C-v C-t org-babel-tangle
C-c C-v f or C-c C-v C-f org-babel-tangle-file
C-c C-v c or C-c C-v C-c org-babel-check-src-block
C-c C-v j or C-c C-v C-j org-babel-insert-header-arg
C-c C-v l or C-c C-v C-l org-babel-load-in-session
C-c C-v i or C-c C-v C-i org-babel-lob-ingest
C-c C-v I or C-c C-v C-I org-babel-view-src-block-info
C-c C-v z or C-c C-v C-z org-babel-switch-to-session-with-code
C-c C-v a or C-c C-v C-a org-babel-sha1-hash
C-c C-v h or C-c C-v C-h org-babel-describe-bindings
C-c C-v x or C-c C-v C-x org-babel-do-key-sequence-in-edit-buffer

Previous: Key bindings and useful functions, Up: Working With Source Code

14.12 Batch execution

It is possible to call functions from the command line. This shell script calls org-babel-tangle on every one of its arguments.

Be sure to adjust the paths to fit your system.

     # -*- mode: shell-script -*-
     # tangle files with org-mode
     # wrap each argument in the code required to call tangle on it
     for i in $@; do
         FILES="$FILES \"$i\""
     emacs -Q --batch \
     --eval "(progn
     (add-to-list 'load-path (expand-file-name \"~/src/org/lisp/\"))
     (add-to-list 'load-path (expand-file-name \"~/src/org/contrib/lisp/\" t))
     (require 'org)(require 'org-exp)(require 'ob)(require 'ob-tangle)
     (mapc (lambda (file)
            (find-file (expand-file-name file \"$DIR\"))
            (kill-buffer)) '($FILES)))" 2>&1 |grep tangled

Next: , Previous: Working With Source Code, Up: Top

15 Miscellaneous

Next: , Previous: Miscellaneous, Up: Miscellaneous

15.1 Completion

Emacs would not be Emacs without completion, and Org mode uses it whenever it makes sense. If you prefer an iswitchb- or ido-like interface for some of the completion prompts, you can specify your preference by setting at most one of the variables org-completion-use-iswitchb org-completion-use-ido.

Org supports in-buffer completion. This type of completion does not make use of the minibuffer. You simply type a few letters into the buffer and use the key to complete text right there.

Complete word at point
  • At the beginning of a headline, complete TODO keywords.
  • After ‘\’, complete TeX symbols supported by the exporter.
  • After ‘*’, complete headlines in the current buffer so that they can be used in search links like ‘[[*find this headline]]’.
  • After ‘:’ in a headline, complete tags. The list of tags is taken from the variable org-tag-alist (possibly set through the ‘#+TAGS’ in-buffer option, see Setting tags), or it is created dynamically from all tags used in the current buffer.
  • After ‘:’ and not in a headline, complete property keys. The list of keys is constructed dynamically from all keys used in the current buffer.
  • After ‘[’, complete link abbreviations (see Link abbreviations).
  • After ‘#+’, complete the special keywords like ‘TYP_TODO’ or ‘OPTIONS’ which set file-specific options for Org mode. When the option keyword is already complete, pressing M-<TAB> again will insert example settings for this keyword.
  • In the line after ‘#+STARTUP: ’, complete startup keywords, i.e., valid keys for this line.
  • Elsewhere, complete dictionary words using Ispell.

Next: , Previous: Completion, Up: Miscellaneous

15.2 Easy Templates

Org mode supports insertion of empty structural elements (like #+BEGIN_SRC and #+END_SRC pairs) with just a few key strokes. This is achieved through a native template expansion mechanism. Note that Emacs has several other template mechanisms which could be used in a similar way, for example yasnippet.

To insert a structural element, type a ‘<’, followed by a template selector and <TAB>. Completion takes effect only when the above keystrokes are typed on a line by itself.

The following template selectors are currently supported.

l #+BEGIN_LaTeX ... #+END_LaTeX
L #+LaTeX:
i #+INDEX: line
I #+INCLUDE: line

For example, on an empty line, typing "<e" and then pressing TAB, will expand into a complete EXAMPLE template.

You can install additional templates by customizing the variable org-structure-template-alist. See the docstring of the variable for additional details.

Next: , Previous: Easy Templates, Up: Miscellaneous

15.3 Speed keys

Single keys can be made to execute commands when the cursor is at the beginning of a headline, i.e., before the first star. Configure the variable org-use-speed-commands to activate this feature. There is a pre-defined list of commands, and you can add more such commands using the variable org-speed-commands-user. Speed keys do not only speed up navigation and other commands, but they also provide an alternative way to execute commands bound to keys that are not or not easily available on a TTY, or on a small mobile device with a limited keyboard.

To see which commands are available, activate the feature and press ? with the cursor at the beginning of a headline.

Next: , Previous: Speed keys, Up: Miscellaneous

15.4 Code evaluation and security issues

Org provides tools to work with the code snippets, including evaluating them.

Running code on your machine always comes with a security risk. Badly written or malicious code can be executed on purpose or by accident. Org has default settings which will only evaluate such code if you give explicit permission to do so, and as a casual user of these features you should leave these precautions intact.

For people who regularly work with such code, the confirmation prompts can become annoying, and you might want to turn them off. This can be done, but you must be aware of the risks that are involved.

Code evaluation can happen under the following circumstances:

Source code blocks
Source code blocks can be evaluated during export, or when pressing C-c C-c in the block. The most important thing to realize here is that Org mode files which contain code snippets are, in a certain sense, like executable files. So you should accept them and load them into Emacs only from trusted sources—just like you would do with a program you install on your computer.

Make sure you know what you are doing before customizing the variables which take off the default security brakes.

— User Option: org-confirm-babel-evaluate

When t (the default), the user is asked before every code block evaluation. When nil, the user is not asked. When set to a function, it is called with two arguments (language and body of the code block) and should return t to ask and nil not to ask.

For example, here is how to execute "ditaa" code (which is considered safe) without asking:

          (defun my-org-confirm-babel-evaluate (lang body)
            (not (string= lang "ditaa")))  ; don't ask for ditaa
          (setq org-confirm-babel-evaluate 'my-org-confirm-babel-evaluate)

Following shell and elisp links
Org has two link types that can directly evaluate code (see External links). These links can be problematic because the code to be evaluated is not visible.
— User Option: org-confirm-shell-link-function

Function to queries user about shell link execution.

— User Option: org-confirm-elisp-link-function

Functions to query user for Emacs Lisp link execution.

Formulas in tables
Formulas in tables (see The spreadsheet) are code that is evaluated either by the calc interpreter, or by the Emacs Lisp interpreter.

Next: , Previous: Code evaluation security, Up: Miscellaneous

15.5 Customization

There are more than 500 variables that can be used to customize Org. For the sake of compactness of the manual, I am not describing the variables here. A structured overview of customization variables is available with M-x org-customize RET. Or select Browse Org Group from the Org->Customization menu. Many settings can also be activated on a per-file basis, by putting special lines into the buffer (see In-buffer settings).

Next: , Previous: Customization, Up: Miscellaneous

15.6 Summary of in-buffer settings

Org mode uses special lines in the buffer to define settings on a per-file basis. These lines start with a ‘#+’ followed by a keyword, a colon, and then individual words defining a setting. Several setting words can be in the same line, but you can also have multiple lines for the keyword. While these settings are described throughout the manual, here is a summary. After changing any of those lines in the buffer, press C-c C-c with the cursor still in the line to activate the changes immediately. Otherwise they become effective only when the file is visited again in a new Emacs session.

#+ARCHIVE: %s_done::
This line sets the archive location for the agenda file. It applies for all subsequent lines until the next ‘#+ARCHIVE’ line, or the end of the file. The first such line also applies to any entries before it. The corresponding variable is org-archive-location.
This line sets the category for the agenda file. The category applies for all subsequent lines until the next ‘#+CATEGORY’ line, or the end of the file. The first such line also applies to any entries before it.
#+COLUMNS: %25ITEM ...
Set the default format for columns view. This format applies when columns view is invoked in locations where no COLUMNS property applies.
#+CONSTANTS: name1=value1 ...
Set file-local values for constants to be used in table formulas. This line sets the local variable org-table-formula-constants-local. The global version of this variable is org-table-formula-constants.
#+FILETAGS: :tag1:tag2:tag3:
Set tags that can be inherited by any entry in the file, including the top-level entries.
Set the file-local set of additional drawers. The corresponding global variable is org-drawers.
#+LINK: linkword replace
These lines (several are allowed) specify link abbreviations. See Link abbreviations. The corresponding variable is org-link-abbrev-alist.
#+PRIORITIES: highest lowest default
This line sets the limits and the default for the priorities. All three must be either letters A–Z or numbers 0–9. The highest priority must have a lower ASCII number than the lowest priority.
#+PROPERTY: Property_Name Value
This line sets a default inheritance value for entries in the current buffer, most useful for specifying the allowed values of a property.
This line defines a file that holds more in-buffer setup. Normally this is entirely ignored. Only when the buffer is parsed for option-setting lines (i.e., when starting Org mode for a file, when pressing C-c C-c in a settings line, or when exporting), then the contents of this file are parsed as if they had been included in the buffer. In particular, the file can be any other Org mode file with internal setup. You can visit the file the cursor is in the line with C-c '.
This line sets options to be used at startup of Org mode, when an Org file is being visited.

The first set of options deals with the initial visibility of the outline tree. The corresponding variable for global default settings is org-startup-folded, with a default value t, which means overview.

          overview         top-level headlines only
          content          all headlines
          showall          no folding of any entries
          showeverything   show even drawer contents

Dynamic virtual indentation is controlled by the variable org-startup-indented173

          indent     start with org-indent-mode turned on
          noindent   start with org-indent-mode turned off

Then there are options for aligning tables upon visiting a file. This is useful in files containing narrowed table columns. The corresponding variable is org-startup-align-all-tables, with a default value nil.

          align      align all tables
          noalign    don't align tables on startup

When visiting a file, inline images can be automatically displayed. The corresponding variable is org-startup-with-inline-images, with a default value nil to avoid delays when visiting a file.

          inlineimages   show inline images
          noinlineimages don't show inline images on startup

When visiting a file, LaTeX fragments can be converted to images automatically. The variable org-startup-with-latex-preview which controls this behavior, is set to nil by default to avoid delays on startup.

          latexpreview   preview LaTeX fragments
          nolatexpreview don't preview LaTeX fragments

Logging the closing and reopening of TODO items and clock intervals can be configured using these options (see variables org-log-done, org-log-note-clock-out and org-log-repeat)

          logdone             record a timestamp when an item is marked DONE
          lognotedone         record timestamp and a note when DONE
          nologdone           don't record when items are marked DONE
          logrepeat           record a time when reinstating a repeating item
          lognoterepeat       record a note when reinstating a repeating item
          nologrepeat         do not record when reinstating repeating item
          lognoteclock-out    record a note when clocking out
          nolognoteclock-out  don't record a note when clocking out
          logreschedule       record a timestamp when scheduling time changes
          lognotereschedule   record a note when scheduling time changes
          nologreschedule     do not record when a scheduling date changes
          logredeadline       record a timestamp when deadline changes
          lognoteredeadline   record a note when deadline changes
          nologredeadline     do not record when a deadline date changes
          logrefile           record a timestamp when refiling
          lognoterefile       record a note when refiling
          nologrefile         do not record when refiling
          logdrawer           store log into drawer
          nologdrawer         store log outside of drawer
          logstatesreversed   reverse the order of states notes
          nologstatesreversed do not reverse the order of states notes

Here are the options for hiding leading stars in outline headings, and for indenting outlines. The corresponding variables are org-hide-leading-stars and org-odd-levels-only, both with a default setting nil (meaning showstars and oddeven).

          hidestars  make all but one of the stars starting a headline invisible.
          showstars  show all stars starting a headline
          indent     virtual indentation according to outline level
          noindent   no virtual indentation according to outline level
          odd        allow only odd outline levels (1,3,...)
          oddeven    allow all outline levels

To turn on custom format overlays over timestamps (variables org-put-time-stamp-overlays and org-time-stamp-overlay-formats), use

          customtime overlay custom time format

The following options influence the table spreadsheet (variable constants-unit-system).

          constcgs   constants.el should use the c-g-s unit system
          constSI    constants.el should use the SI unit system

To influence footnote settings, use the following keywords. The corresponding variables are org-footnote-define-inline, org-footnote-auto-label, and org-footnote-auto-adjust.

          fninline    define footnotes inline
          fnnoinline  define footnotes in separate section
          fnlocal     define footnotes near first reference, but not inline
          fnprompt    prompt for footnote labels
          fnauto      create [fn:1]-like labels automatically (default)
          fnconfirm   offer automatic label for editing or confirmation
          fnplain     create [1]-like labels automatically
          fnadjust    automatically renumber and sort footnotes
          nofnadjust  do not renumber and sort automatically

To hide blocks on startup, use these keywords. The corresponding variable is org-hide-block-startup.

          hideblocks   Hide all begin/end blocks on startup
          nohideblocks Do not hide blocks on startup

The display of entities as UTF-8 characters is governed by the variable org-pretty-entities and the keywords

          entitiespretty  Show entities as UTF-8 characters where possible
          entitiesplain   Leave entities plain

#+TAGS: TAG1(c1) TAG2(c2)
These lines (several such lines are allowed) specify the valid tags in this file, and (potentially) the corresponding fast tag selection keys. The corresponding variable is org-tag-alist.
This line contains the formulas for the table directly above the line.

Table can have multiple lines containing ‘#+TBLFM:’. Note that only the first line of ‘#+TBLFM:’ will be applied when you recalculate the table. For more details see Using multiple #+TBLFM lines in Editing and debugging formulas.

These lines provide settings for exporting files. For more details see Export settings.
These lines set the TODO keywords and their interpretation in the current file. The corresponding variable is org-todo-keywords.

Next: , Previous: In-buffer settings, Up: Miscellaneous

15.7 The very busy C-c C-c key

The key C-c C-c has many purposes in Org, which are all mentioned scattered throughout this manual. One specific function of this key is to add tags to a headline (see Tags). In many other circumstances it means something like “Hey Org, look here and update according to what you see here”. Here is a summary of what this means in different contexts.

Next: , Previous: The very busy C-c C-c key, Up: Miscellaneous

15.8 A cleaner outline view

Some people find it noisy and distracting that the Org headlines start with a potentially large number of stars, and that text below the headlines is not indented. While this is no problem when writing a book-like document where the outline headings are really section headings, in a more list-oriented outline, indented structure is a lot cleaner:

     * Top level headline             |    * Top level headline
     ** Second level                  |      * Second level
     *** 3rd level                    |        * 3rd level
     some text                        |          some text
     *** 3rd level                    |        * 3rd level
     more text                        |          more text
     * Another top level headline     |    * Another top level headline

If you are using at least Emacs 23.2174 and version 6.29 of Org, this kind of view can be achieved dynamically at display time using org-indent-mode. In this minor mode, all lines are prefixed for display with the necessary amount of space175. Also headlines are prefixed with additional stars, so that the amount of indentation shifts by two176 spaces per level. All headline stars but the last one are made invisible using the org-hide face177; see below under ‘2.’ for more information on how this works. You can turn on org-indent-mode for all files by customizing the variable org-startup-indented, or you can turn it on for individual files using

     #+STARTUP: indent

If you want a similar effect in an earlier version of Emacs and/or Org, or if you want the indentation to be hard space characters so that the plain text file looks as similar as possible to the Emacs display, Org supports you in the following way:

  1. Indentation of text below headlines
    You may indent text below each headline to make the left boundary line up with the headline, like
              *** 3rd level
                  more text, now indented

    Org supports this with paragraph filling, line wrapping, and structure editing178, preserving or adapting the indentation as appropriate.

  2. Hiding leading stars
    You can modify the display in such a way that all leading stars become invisible. To do this in a global way, configure the variable org-hide-leading-stars or change this on a per-file basis with
              #+STARTUP: hidestars
              #+STARTUP: showstars

    With hidden stars, the tree becomes:

              * Top level headline
               * Second level
                * 3rd level

    The leading stars are not truly replaced by whitespace, they are only fontified with the face org-hide that uses the background color as font color. If you are not using either white or black background, you may have to customize this face to get the wanted effect. Another possibility is to set this font such that the extra stars are almost invisible, for example using the color grey90 on a white background.

  3. Things become cleaner still if you skip all the even levels and use only odd levels 1, 3, 5..., effectively adding two stars to go from one outline level to the next179. In this way we get the outline view shown at the beginning of this section. In order to make the structure editing and export commands handle this convention correctly, configure the variable org-odd-levels-only, or set this on a per-file basis with one of the following lines:
              #+STARTUP: odd
              #+STARTUP: oddeven

    You can convert an Org file from single-star-per-level to the double-star-per-level convention with M-x org-convert-to-odd-levels RET in that file. The reverse operation is M-x org-convert-to-oddeven-levels.

Next: , Previous: Clean view, Up: Miscellaneous

15.9 Using Org on a tty

Because Org contains a large number of commands, by default many of Org's core commands are bound to keys that are generally not accessible on a tty, such as the cursor keys (<left>, <right>, <up>, <down>), <TAB> and <RET>, in particular when used together with modifiers like <Meta> and/or <Shift>. To access these commands on a tty when special keys are unavailable, the following alternative bindings can be used. The tty bindings below will likely be more cumbersome; you may find for some of the bindings below that a customized workaround suits you better. For example, changing a timestamp is really only fun with S-<cursor> keys, whereas on a tty you would rather use C-c . to re-insert the timestamp.

Default Alternative 1 Speed key Alternative 2
S-<TAB> C-u <TAB> C
M-<left> C-c C-x l l <Esc> <left>
M-S-<left> C-c C-x L L
M-<right> C-c C-x r r <Esc> <right>
M-S-<right> C-c C-x R R
M-<up> C-c C-x u <Esc> <up>
M-S-<up> C-c C-x U U
M-<down> C-c C-x d <Esc> <down>
M-S-<down> C-c C-x D D
S-<RET> C-c C-x c
M-<RET> C-c C-x m <Esc> <RET>
M-S-<RET> C-c C-x M
S-<left> C-c <left>
S-<right> C-c <right>
S-<up> C-c <up>
S-<down> C-c <down>
C-S-<left> C-c C-x <left>
C-S-<right> C-c C-x <right>

Next: , Previous: TTY keys, Up: Miscellaneous

15.10 Interaction with other packages

Org lives in the world of GNU Emacs and interacts in various ways with other code out there.

Next: , Previous: Interaction, Up: Interaction

15.10.1 Packages that Org cooperates with

calc.el by Dave Gillespie
Org uses the Calc package for implementing spreadsheet functionality in its tables (see The spreadsheet). Org checks for the availability of Calc by looking for the function calc-eval which will have been autoloaded during setup if Calc has been installed properly. As of Emacs 22, Calc is part of the Emacs distribution. Another possibility for interaction between the two packages is using Calc for embedded calculations. See Embedded Mode.
constants.el by Carsten Dominik
In a table formula (see The spreadsheet), it is possible to use names for natural constants or units. Instead of defining your own constants in the variable org-table-formula-constants, install the constants package which defines a large number of constants and units, and lets you use unit prefixes like ‘M’ for ‘Mega’, etc. You will need version 2.0 of this package, available at Org checks for the function constants-get, which has to be autoloaded in your setup. See the installation instructions in the file constants.el.
cdlatex.el by Carsten Dominik
Org mode can make use of the CDLaTeX package to efficiently enter LaTeX fragments into Org files. See CDLaTeX mode.
imenu.el by Ake Stenhoff and Lars Lindberg
Imenu allows menu access to an index of items in a file. Org mode supports Imenu—all you need to do to get the index is the following:
          (add-hook 'org-mode-hook
                    (lambda () (imenu-add-to-menubar "Imenu")))

By default the index is two levels deep—you can modify the depth using the option org-imenu-depth.

remember.el by John Wiegley
Org used to use this package for capture, but no longer does.
speedbar.el by Eric M. Ludlam
Speedbar is a package that creates a special frame displaying files and index items in files. Org mode supports Speedbar and allows you to drill into Org files directly from the Speedbar. It also allows you to restrict the scope of agenda commands to a file or a subtree by using the command < in the Speedbar frame.
table.el by Takaaki Ota
Complex ASCII tables with automatic line wrapping, column- and row-spanning, and alignment can be created using the Emacs table package by Takaaki Ota (, and also part of Emacs 22). Org mode will recognize these tables and export them properly. Because of interference with other Org mode functionality, you unfortunately cannot edit these tables directly in the buffer. Instead, you need to use the command C-c ' to edit them, similar to source code snippets.
C-c '     (org-edit-special)
Edit a table.el table. Works when the cursor is in a table.el table.
C-c ~     (org-table-create-with-table.el)
Insert a table.el table. If there is already a table at point, this command converts it between the table.el format and the Org mode format. See the documentation string of the command org-convert-table for the restrictions under which this is possible.
table.el is part of Emacs since Emacs 22.
footnote.el by Steven L. Baur
Org mode recognizes numerical footnotes as provided by this package. However, Org mode also has its own footnote support (see Footnotes), which makes using footnote.el unnecessary.

Previous: Cooperation, Up: Interaction

15.10.2 Packages that lead to conflicts with Org mode

In Emacs 23, shift-selection-mode is on by default, meaning that cursor motions combined with the shift key should start or enlarge regions. This conflicts with the use of S-<cursor> commands in Org to change timestamps, TODO keywords, priorities, and item bullet types if the cursor is at such a location. By default, S-<cursor> commands outside special contexts don't do anything, but you can customize the variable org-support-shift-select. Org mode then tries to accommodate shift selection by (i) using it outside of the special contexts where special commands apply, and by (ii) extending an existing active region even if the cursor moves across a special context.
CUA.el by Kim. F. Storm
Key bindings in Org conflict with the S-<cursor> keys used by CUA mode (as well as pc-select-mode and s-region-mode) to select and extend the region. In fact, Emacs 23 has this built-in in the form of shift-selection-mode, see previous paragraph. If you are using Emacs 23, you probably don't want to use another package for this purpose. However, if you prefer to leave these keys to a different package while working in Org mode, configure the variable org-replace-disputed-keys. When set, Org will move the following key bindings in Org files, and in the agenda buffer (but not during date selection).
          S-UP      ⇒  M-p             S-DOWN     ⇒  M-n
          S-LEFT    ⇒  M--             S-RIGHT    ⇒  M-+
          C-S-LEFT  ⇒  M-S--           C-S-RIGHT  ⇒  M-S-+

Yes, these are unfortunately more difficult to remember. If you want to have other replacement keys, look at the variable org-disputed-keys.

ecomplete.el by Lars Magne Ingebrigtsen
Ecomplete provides “electric” address completion in address header lines in message buffers. Sadly Orgtbl mode cuts ecompletes power supply: No completion happens when Orgtbl mode is enabled in message buffers while entering text in address header lines. If one wants to use ecomplete one should not follow the advice to automagically turn on Orgtbl mode in message buffers (see Orgtbl mode), but instead—after filling in the message headers—turn on Orgtbl mode manually when needed in the messages body.
filladapt.el by Kyle Jones
Org mode tries to do the right thing when filling paragraphs, list items and other elements. Many users reported they had problems using both filladapt.el and Org mode, so a safe thing to do is to disable it like this:
          (add-hook 'org-mode-hook 'turn-off-filladapt-mode)

The way Org mode binds the <TAB> key (binding to [tab] instead of "\t") overrules YASnippet's access to this key. The following code fixed this problem:
          (add-hook 'org-mode-hook
                    (lambda ()
                      (org-set-local 'yas/trigger-key [tab])
                      (define-key yas/keymap [tab] 'yas/next-field-or-maybe-expand)))

The latest version of yasnippet doesn't play well with Org mode. If the above code does not fix the conflict, start by defining the following function:

          (defun yas/org-very-safe-expand ()
            (let ((yas/fallback-behavior 'return-nil)) (yas/expand)))

Then, tell Org mode what to do with the new function:

          (add-hook 'org-mode-hook
                    (lambda ()
                      (make-variable-buffer-local 'yas/trigger-key)
                      (setq yas/trigger-key [tab])
                      (add-to-list 'org-tab-first-hook 'yas/org-very-safe-expand)
                      (define-key yas/keymap [tab] 'yas/next-field)))

windmove.el by Hovav Shacham
This package also uses the S-<cursor> keys, so everything written in the paragraph above about CUA mode also applies here. If you want make the windmove function active in locations where Org mode does not have special functionality on S-<cursor>, add this to your configuration:
          ;; Make windmove work in org-mode:
          (add-hook 'org-shiftup-final-hook 'windmove-up)
          (add-hook 'org-shiftleft-final-hook 'windmove-left)
          (add-hook 'org-shiftdown-final-hook 'windmove-down)
          (add-hook 'org-shiftright-final-hook 'windmove-right)

viper.el by Michael Kifer
Viper uses C-c / and therefore makes this key not access the corresponding Org mode command org-sparse-tree. You need to find another key for this command, or override the key in viper-vi-global-user-map with
          (define-key viper-vi-global-user-map "C-c /" 'org-sparse-tree)

Previous: Interaction, Up: Miscellaneous

15.11 org-crypt.el

Org-crypt will encrypt the text of an entry, but not the headline, or properties. Org-crypt uses the Emacs EasyPG library to encrypt and decrypt files.

Any text below a headline that has a ‘:crypt:’ tag will be automatically be encrypted when the file is saved. If you want to use a different tag just customize the org-crypt-tag-matcher setting.

To use org-crypt it is suggested that you have the following in your .emacs:

     (require 'org-crypt)
     (setq org-tags-exclude-from-inheritance (quote ("crypt")))
     (setq org-crypt-key nil)
       ;; GPG key to use for encryption
       ;; Either the Key ID or set to nil to use symmetric encryption.
     (setq auto-save-default nil)
       ;; Auto-saving does not cooperate with org-crypt.el: so you need
       ;; to turn it off if you plan to use org-crypt.el quite often.
       ;; Otherwise, you'll get an (annoying) message each time you
       ;; start Org.
       ;; To turn it off only locally, you can insert this:
       ;; # -*- buffer-auto-save-file-name: nil; -*-

Excluding the crypt tag from inheritance prevents already encrypted text being encrypted again.

Next: , Previous: Miscellaneous, Up: Top

Appendix A Hacking

This appendix covers some aspects where users can extend the functionality of Org.

Next: , Previous: Hacking, Up: Hacking

A.1 Hooks

Org has a large number of hook variables that can be used to add functionality. This appendix about hacking is going to illustrate the use of some of them. A complete list of all hooks with documentation is maintained by the Worg project and can be found at

Next: , Previous: Hooks, Up: Hacking

A.2 Add-on packages

A large number of add-on packages have been written by various authors.

These packages are not part of Emacs, but they are distributed as contributed packages with the separate release available at See the contrib/README file in the source code directory for a list of contributed files. You may also find some more information on the Worg page:

Next: , Previous: Add-on packages, Up: Hacking

A.3 Adding hyperlink types

Org has a large number of hyperlink types built-in (see Hyperlinks). If you would like to add new link types, Org provides an interface for doing so. Let's look at an example file, org-man.el, that will add support for creating links like ‘[[man:printf][The printf manpage]]’ to show Unix manual pages inside Emacs:

     ;;; org-man.el - Support for links to manpages in Org
     (require 'org)
     (org-add-link-type "man" 'org-man-open)
     (add-hook 'org-store-link-functions 'org-man-store-link)
     (defcustom org-man-command 'man
       "The Emacs command to be used to display a man page."
       :group 'org-link
       :type '(choice (const man) (const woman)))
     (defun org-man-open (path)
       "Visit the manpage on PATH.
     PATH should be a topic that can be thrown at the man command."
       (funcall org-man-command path))
     (defun org-man-store-link ()
       "Store a link to a manpage."
       (when (memq major-mode '(Man-mode woman-mode))
         ;; This is a man page, we do make this link
         (let* ((page (org-man-get-page-name))
                (link (concat "man:" page))
                (description (format "Manpage for %s" page)))
            :type "man"
            :link link
            :description description))))
     (defun org-man-get-page-name ()
       "Extract the page name from the buffer name."
       ;; This works for both `Man-mode' and `woman-mode'.
       (if (string-match " \\(\\S-+\\)\\*" (buffer-name))
           (match-string 1 (buffer-name))
         (error "Cannot create link to this man page")))
     (provide 'org-man)
     ;;; org-man.el ends here

You would activate this new link type in .emacs with

     (require 'org-man)

Let's go through the file and see what it does.

  1. It does (require 'org) to make sure that org.el has been loaded.
  2. The next line calls org-add-link-type to define a new link type with prefix ‘man’. The call also contains the name of a function that will be called to follow such a link.
  3. The next line adds a function to org-store-link-functions, in order to allow the command C-c l to record a useful link in a buffer displaying a man page.

The rest of the file defines the necessary variables and functions. First there is a customization variable that determines which Emacs command should be used to display man pages. There are two options, man and woman. Then the function to follow a link is defined. It gets the link path as an argument—in this case the link path is just a topic for the manual command. The function calls the value of org-man-command to display the man page.

Finally the function org-man-store-link is defined. When you try to store a link with C-c l, this function will be called to try to make a link. The function must first decide if it is supposed to create the link for this buffer type; we do this by checking the value of the variable major-mode. If not, the function must exit and return the value nil. If yes, the link is created by getting the manual topic from the buffer name and prefixing it with the string ‘man:’. Then it must call the command org-store-link-props and set the :type and :link properties. Optionally you can also set the :description property to provide a default for the link description when the link is later inserted into an Org buffer with C-c C-l.

When it makes sense for your new link type, you may also define a function org-PREFIX-complete-link that implements special (e.g., completion) support for inserting such a link with C-c C-l. Such a function should not accept any arguments, and return the full link with prefix.

Next: , Previous: Adding hyperlink types, Up: Hacking

A.4 Adding export back-ends

Org 8.0 comes with a completely rewritten export engine which makes it easy to write new export back-ends, either from scratch, or from deriving them from existing ones.

Your two entry points are respectively org-export-define-backend and org-export-define-derived-backend. To grok these functions, you should first have a look at ox-latex.el (for how to define a new back-end from scratch) and ox-beamer.el (for how to derive a new back-end from an existing one.

When creating a new back-end from scratch, the basic idea is to set the name of the back-end (as a symbol) and an an alist of elements and export functions. On top of this, you will need to set additional keywords like :menu-entry (to display the back-end in the export dispatcher), :export-block (to specify what blocks should not be exported by this back-end), and :options-alist (to let the user set export options that are specific to this back-end.)

Deriving a new back-end is similar, except that you need to set :translate-alist to an alist of export functions that should be used instead of the parent back-end functions.

For a complete reference documentation, see the Org Export Reference on Worg.

Next: , Previous: Adding export back-ends, Up: Hacking

A.5 Context-sensitive commands

Org has several commands that act differently depending on context. The most important example is the C-c C-c (see The very busy C-c C-c key). Also the M-cursor and M-S-cursor keys have this property.

Add-ons can tap into this functionality by providing a function that detects special context for that add-on and executes functionality appropriate for the context. Here is an example from Dan Davison's org-R.el which allows you to evaluate commands based on the R programming language 180. For this package, special contexts are lines that start with #+R: or #+RR:.

     (defun org-R-apply-maybe ()
       "Detect if this is context for org-R and execute R commands."
       (if (save-excursion
             (beginning-of-line 1)
             (looking-at "#\\+RR?:"))
           (progn (call-interactively 'org-R-apply)
                  t) ;; to signal that we took action
         nil)) ;; to signal that we did not
     (add-hook 'org-ctrl-c-ctrl-c-hook 'org-R-apply-maybe)

The function first checks if the cursor is in such a line. If that is the case, org-R-apply is called and the function returns t to signal that action was taken, and C-c C-c will stop looking for other contexts. If the function finds it should do nothing locally, it returns nil so that other, similar functions can have a try.

Next: , Previous: Context-sensitive commands, Up: Hacking

A.6 Tables and lists in arbitrary syntax

Since Orgtbl mode can be used as a minor mode in arbitrary buffers, a frequent feature request has been to make it work with native tables in specific languages, for example LaTeX. However, this is extremely hard to do in a general way, would lead to a customization nightmare, and would take away much of the simplicity of the Orgtbl mode table editor.

This appendix describes a different approach. We keep the Orgtbl mode table in its native format (the source table), and use a custom function to translate the table to the correct syntax, and to install it in the right location (the target table). This puts the burden of writing conversion functions on the user, but it allows for a very flexible system.

Bastien added the ability to do the same with lists, in Orgstruct mode. You can use Org's facilities to edit and structure lists by turning orgstruct-mode on, then locally exporting such lists in another format (HTML, LaTeX or Texinfo.)

Next: , Previous: Tables in arbitrary syntax, Up: Tables in arbitrary syntax

A.6.1 Radio tables

To define the location of the target table, you first need to create two lines that are comments in the current mode, but contain magic words BEGIN/END RECEIVE ORGTBL for Orgtbl mode to find. Orgtbl mode will insert the translated table between these lines, replacing whatever was there before. For example in C mode where comments are between /* ... */:

     /* BEGIN RECEIVE ORGTBL table_name */
     /* END RECEIVE ORGTBL table_name */

Just above the source table, we put a special line that tells Orgtbl mode how to translate this table and where to install it. For example:

     #+ORGTBL: SEND table_name translation_function arguments...

table_name is the reference name for the table that is also used in the receiver lines. translation_function is the Lisp function that does the translation. Furthermore, the line can contain a list of arguments (alternating key and value) at the end. The arguments will be passed as a property list to the translation function for interpretation. A few standard parameters are already recognized and acted upon before the translation function is called:

:skip N
Skip the first N lines of the table. Hlines do count as separate lines for this parameter!
:skipcols (n1 n2 ...)
List of columns that should be skipped. If the table has a column with calculation marks, that column is automatically discarded as well. Please note that the translator function sees the table after the removal of these columns, the function never knows that there have been additional columns.
:no-escape t
When non-nil, do not escape special characters &%#_^ when exporting the table. The default value is nil.

The one problem remaining is how to keep the source table in the buffer without disturbing the normal workings of the file, for example during compilation of a C file or processing of a LaTeX file. There are a number of different solutions:

Next: , Previous: Radio tables, Up: Tables in arbitrary syntax

A.6.2 A LaTeX example of radio tables

The best way to wrap the source table in LaTeX is to use the comment environment provided by comment.sty. It has to be activated by placing \usepackage{comment} into the document header. Orgtbl mode can insert a radio table skeleton181 with the command M-x orgtbl-insert-radio-table RET. You will be prompted for a table name, let's say we use ‘salesfigures’. You will then get the following template:

     % BEGIN RECEIVE ORGTBL salesfigures
     % END RECEIVE ORGTBL salesfigures
     #+ORGTBL: SEND salesfigures orgtbl-to-latex
     | | |

The #+ORGTBL: SEND line tells Orgtbl mode to use the function orgtbl-to-latex to convert the table into LaTeX and to put it into the receiver location with name salesfigures. You may now fill in the table—feel free to use the spreadsheet features182:

     % BEGIN RECEIVE ORGTBL salesfigures
     % END RECEIVE ORGTBL salesfigures
     #+ORGTBL: SEND salesfigures orgtbl-to-latex
     | Month | Days | Nr sold | per day |
     | Jan   |   23 |      55 |     2.4 |
     | Feb   |   21 |      16 |     0.8 |
     | March |   22 |     278 |    12.6 |
     #+TBLFM: $4=$3/$2;%.1f
     % $ (optional extra dollar to keep font-lock happy, see footnote)

When you are done, press C-c C-c in the table to get the converted table inserted between the two marker lines.

Now let's assume you want to make the table header by hand, because you want to control how columns are aligned, etc. In this case we make sure that the table translator skips the first 2 lines of the source table, and tell the command to work as a splice, i.e., to not produce header and footer commands of the target table:

     Month & \multicolumn{1}{c}{Days} & Nr.\ sold & per day\\
     % BEGIN RECEIVE ORGTBL salesfigures
     % END RECEIVE ORGTBL salesfigures
     #+ORGTBL: SEND salesfigures orgtbl-to-latex :splice t :skip 2
     | Month | Days | Nr sold | per day |
     | Jan   |   23 |      55 |     2.4 |
     | Feb   |   21 |      16 |     0.8 |
     | March |   22 |     278 |    12.6 |
     #+TBLFM: $4=$3/$2;%.1f

The LaTeX translator function orgtbl-to-latex is already part of Orgtbl mode. It uses a tabular environment to typeset the table and marks horizontal lines with \hline. Furthermore, it interprets the following parameters (see also see Translator functions):

:splice nil/t
When set to t, return only table body lines, don't wrap them into a tabular environment. Default is nil.
:fmt fmt
A format to be used to wrap each field, it should contain %s for the original field value. For example, to wrap each field value in dollars, you could use :fmt "$%s$". This may also be a property list with column numbers and formats, for example :fmt (2 "$%s$" 4 "%s\\%%"). A function of one argument can be used in place of the strings; the function must return a formatted string.
:efmt efmt
Use this format to print numbers with exponentials. The format should have %s twice for inserting mantissa and exponent, for example "%s\\times10^{%s}". The default is "%s\\,(%s)". This may also be a property list with column numbers and formats, for example :efmt (2 "$%s\\times10^{%s}$" 4 "$%s\\cdot10^{%s}$"). After efmt has been applied to a value, fmt will also be applied. Similar to fmt, functions of two arguments can be supplied instead of strings.

Next: , Previous: A LaTeX example, Up: Tables in arbitrary syntax

A.6.3 Translator functions

Orgtbl mode has several translator functions built-in: orgtbl-to-csv (comma-separated values), orgtbl-to-tsv (TAB-separated values) orgtbl-to-latex, orgtbl-to-html, and orgtbl-to-texinfo. Except for orgtbl-to-html183, these all use a generic translator, orgtbl-to-generic. For example, orgtbl-to-latex itself is a very short function that computes the column definitions for the tabular environment, defines a few field and line separators and then hands processing over to the generic translator. Here is the entire code:

     (defun orgtbl-to-latex (table params)
       "Convert the Orgtbl mode TABLE to LaTeX."
       (let* ((alignment (mapconcat (lambda (x) (if x "r" "l"))
                                    org-table-last-alignment ""))
                :tstart (concat "\\begin{tabular}{" alignment "}")
                :tend "\\end{tabular}"
                :lstart "" :lend " \\\\" :sep " & "
                :efmt "%s\\,(%s)" :hline "\\hline")))
         (orgtbl-to-generic table (org-combine-plists params2 params))))

As you can see, the properties passed into the function (variable PARAMS) are combined with the ones newly defined in the function (variable PARAMS2). The ones passed into the function (i.e., the ones set by the ‘ORGTBL SEND’ line) take precedence. So if you would like to use the LaTeX translator, but wanted the line endings to be ‘\\[2mm]’ instead of the default ‘\\’, you could just overrule the default with

     #+ORGTBL: SEND test orgtbl-to-latex :lend " \\\\[2mm]"

For a new language, you can either write your own converter function in analogy with the LaTeX translator, or you can use the generic function directly. For example, if you have a language where a table is started with ‘!BTBL!’, ended with ‘!ETBL!’, and where table lines are started with ‘!BL!’, ended with ‘!EL!’, and where the field separator is a TAB, you could call the generic translator like this (on a single line!):

     #+ORGTBL: SEND test orgtbl-to-generic :tstart "!BTBL!" :tend "!ETBL!"
                                   :lstart "!BL! " :lend " !EL!" :sep "\t"

Please check the documentation string of the function orgtbl-to-generic for a full list of parameters understood by that function, and remember that you can pass each of them into orgtbl-to-latex, orgtbl-to-texinfo, and any other function using the generic function.

Of course you can also write a completely new function doing complicated things the generic translator cannot do. A translator function takes two arguments. The first argument is the table, a list of lines, each line either the symbol hline or a list of fields. The second argument is the property list containing all parameters specified in the ‘#+ORGTBL: SEND’ line. The function must return a single string containing the formatted table. If you write a generally useful translator, please post it on so that others can benefit from your work.

Previous: Translator functions, Up: Tables in arbitrary syntax

A.6.4 Radio lists

Sending and receiving radio lists works exactly the same way as sending and receiving radio tables (see Radio tables). As for radio tables, you can insert radio list templates in HTML, LaTeX and Texinfo modes by calling org-list-insert-radio-list.

Here are the differences with radio tables:

Here is a LaTeX example. Let's say that you have this in your LaTeX file:

     % END RECEIVE ORGLST to-buy
     #+ORGLST: SEND to-buy org-list-to-latex
     - a new house
     - a new computer
       + a new keyboard
       + a new mouse
     - a new life

Pressing C-c C-c on a new house and will insert the converted LaTeX list between the two marker lines.

Next: , Previous: Tables in arbitrary syntax, Up: Hacking

A.7 Dynamic blocks

Org documents can contain dynamic blocks. These are specially marked regions that are updated by some user-written function. A good example for such a block is the clock table inserted by the command C-c C-x C-r (see Clocking work time).

Dynamic blocks are enclosed by a BEGIN-END structure that assigns a name to the block and can also specify parameters for the function producing the content of the block.

     #+BEGIN: myblock :parameter1 value1 :parameter2 value2 ...

Dynamic blocks are updated with the following commands

C-c C-x C-u     (org-dblock-update)
Update dynamic block at point.
C-u C-c C-x C-u
Update all dynamic blocks in the current file.

Updating a dynamic block means to remove all the text between BEGIN and END, parse the BEGIN line for parameters and then call the specific writer function for this block to insert the new content. If you want to use the original content in the writer function, you can use the extra parameter :content.

For a block with name myblock, the writer function is org-dblock-write:myblock with as only parameter a property list with the parameters given in the begin line. Here is a trivial example of a block that keeps track of when the block update function was last run:

     #+BEGIN: block-update-time :format "on %m/%d/%Y at %H:%M"

The corresponding block writer function could look like this:

     (defun org-dblock-write:block-update-time (params)
       (let ((fmt (or (plist-get params :format) "%d. %m. %Y")))
         (insert "Last block update at: "
                 (format-time-string fmt))))

If you want to make sure that all dynamic blocks are always up-to-date, you could add the function org-update-all-dblocks to a hook, for example before-save-hook. org-update-all-dblocks is written in a way such that it does nothing in buffers that are not in org-mode.

You can narrow the current buffer to the current dynamic block (like any other block) with org-narrow-to-block.

Next: , Previous: Dynamic blocks, Up: Hacking

A.8 Special agenda views

Org provides a special hook that can be used to narrow down the selection made by these agenda views: agenda, agenda*184, todo, alltodo, tags, tags-todo, tags-tree. You may specify a function that is used at each match to verify if the match should indeed be part of the agenda view, and if not, how much should be skipped. You can specify a global condition that will be applied to all agenda views, this condition would be stored in the variable org-agenda-skip-function-global. More commonly, such a definition is applied only to specific custom searches, using org-agenda-skip-function.

Let's say you want to produce a list of projects that contain a WAITING tag anywhere in the project tree. Let's further assume that you have marked all tree headings that define a project with the TODO keyword PROJECT. In this case you would run a TODO search for the keyword PROJECT, but skip the match unless there is a WAITING tag anywhere in the subtree belonging to the project line.

To achieve this, you must write a function that searches the subtree for the tag. If the tag is found, the function must return nil to indicate that this match should not be skipped. If there is no such tag, return the location of the end of the subtree, to indicate that search should continue from there.

     (defun my-skip-unless-waiting ()
       "Skip trees that are not waiting"
       (let ((subtree-end (save-excursion (org-end-of-subtree t))))
         (if (re-search-forward ":waiting:" subtree-end t)
             nil          ; tag found, do not skip
           subtree-end))) ; tag not found, continue after end of subtree

Now you may use this function in an agenda custom command, for example like this:

      '("b" todo "PROJECT"
        ((org-agenda-skip-function 'my-skip-unless-waiting)
         (org-agenda-overriding-header "Projects waiting for something: "))))

Note that this also binds org-agenda-overriding-header to get a meaningful header in the agenda view.

A general way to create custom searches is to base them on a search for entries with a certain level limit. If you want to study all entries with your custom search function, simply do a search for ‘LEVEL>0185, and then use org-agenda-skip-function to select the entries you really want to have.

You may also put a Lisp form into org-agenda-skip-function. In particular, you may use the functions org-agenda-skip-entry-if and org-agenda-skip-subtree-if in this form, for example:

(org-agenda-skip-entry-if 'scheduled)
Skip current entry if it has been scheduled.
(org-agenda-skip-entry-if 'notscheduled)
Skip current entry if it has not been scheduled.
(org-agenda-skip-entry-if 'deadline)
Skip current entry if it has a deadline.
(org-agenda-skip-entry-if 'scheduled 'deadline)
Skip current entry if it has a deadline, or if it is scheduled.
(org-agenda-skip-entry-if 'todo '(