ESS version 24.01.1
by A.J. Rossini,
R.M. Heiberger,
K. Hornik,
M. Maechler,
R.A. Sparapani,
S.J. Eglen,
S.P. Luque,
H. Redestig,
V. Spinu,
L. Henry,
and J.A. Branham.
Emacs Speaks Statistics (ESS) provides an intelligent, consistent interface between the user and the software. ESS interfaces with R, SAS, S-PLUS, BUGS/JAGS and other statistical analysis packages on GNU/Linux, other Unix-like systems such as macOS, and Microsoft Windows. ESS is itself a package within the Emacs text editor and uses Emacs features to streamline the creation and use of statistical software. ESS knows the syntax and grammar of statistical analysis packages and provides consistent display and editing features based on that knowledge. ESS assists in interactive and batch execution of statements written in these statistical analysis languages.
The source and documentation of ESS is free software. You can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version.
ESS is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License in the file COPYING in the same directory as this file for more details.
ESS provides a generic interface, through Emacs, to statistical packages. It currently supports R (and the rest of the S family), SAS, BUGS/JAGS, Stata, and Julia with the level of support roughly in that order.
Throughout this manual, Emacs refers to GNU Emacs by the Free Software Foundation. Although previous versions of ESS supported other Emacsen, current versions only support GNU Emacs.
There are two main ways of interacting with ESS: through “regular”
modes or “inferior” modes. Regular modes act like normal Emacs major
modes. ESS major modes are displayed in the mode-line in the format
ESS[dialect], where dialect can take values such as
R, SAS, or S.
ESS also provides easy access to an “inferior process,” which is an Emacs buffer associated with a running process. This can be an R session, for example. These inferior processes are referred
to as inferior ESS (iESS), and are shown in the modeline by
iESS [dialect].
Currently, the documentation contains many references to ‘R’ where actually any supported (statistics) language is meant, i.e., ‘R’ could also mean ‘S’ or ‘SAS.’
For exclusively interactive users of R, ESS provides a number of features to make life easier. There is an easy to use command history mechanism, including a quick prefix-search history. To reduce typing, command-line completion is provided for all R objects and “hot keys” are provided for common R function calls. Help files are easily accessible, and a paging mechanism is provided to view them. Finally, an incidental (but very useful) side-effect of ESS is that a transcript of your session is kept for later saving or editing.
No special knowledge of Emacs is necessary when using R interactively under ESS.
For those that use R in the typical edit–test–revise cycle when writing R functions, ESS provides for editing of R functions in Emacs buffers. Unlike the typical use of R where the editor is restarted every time an object is edited, ESS uses the current Emacs session for editing. In practical terms, this means that you can edit more than one function at once, and that the ESS process is still available for use while editing. Error checking is performed on functions loaded back into R, and a mechanism to jump directly to the error is provided. ESS also provides for maintaining text versions of your R functions in specified source directories.
Statistical packages are powerful software systems for manipulating and analyzing data, but their user interfaces often leave something to be desired: they offer weak editor functionality and they differ among themselves so markedly that you have to re-learn how to do those things for each package. ESS is a package which is designed to make editing and interacting with statistical packages more uniform, user-friendly and give you the power of Emacs as well.
Additionally, both Emacs and ESS (and R) are free software designed to give users full control over their computer. For more on what this means, visit https://www.gnu.org/philosophy/free-sw.html.
For source code buffers, ESS offers several features:
ESS also provides features that make it easier to interact with inferior ESS (iESS) process (a connection between your buffer and the statistical package which is waiting for you to input commands). These include:
objects(), and search(). See Hot keys for common commands.
Finally, ESS provides features for re-submitting commands from saved transcript files, including:
Changes and New Features in development version:
Changes and New Features in 24.01.1:
ess-request-a-process change
Changes and New Features in 24.01.0:
ess-request-a-process obeys ess-gen-proc-buffer-name-function,
thanks to Ihor Radchenko.
Changes and New Features in 19.04 (unreleased):
browser() in
an environment that doesn’t inherit from the search path.
options(width = ) is now set on startup based
on the width of the inferior window.
ess-command now runs R code in a sandboxed environment.
Use .ess.environment() to inspect the current environment.
ess-elisp-trace-mode minor mode.
Toggle it to start or stop tracing all ess-prefixed functions
with trace-function. Tracing is useful for debugging background
ESS behaviour.
ess-get-help-aliases-list now caches the aliases
on the R side. This should speed up help lookup when the search path has
changed and the aliases are read again.
ess-command now uses a default timeout of 30 seconds.
It should normally be avoided with long-running tasks because it causes
Emacs to block while the command is running. If the timeout is reached,
an error is thrown. An interrupt is also sent to the process in case of early
exit.
This is a behaviour change: you will now have to explicitly opt in
blocking the whole Emacs UI for more than 30 seconds by supplying a
larger timeout (use most-positive-fixnum for infinity).
ess-wait-for-process now returns nil if a timeout is
reached.
ess-get-words-from-vector gains a timeout argument.
ess-command is now more robust and resilient to hangs
and custom prompts (#1043). It also strips continuation prompts (+
prompts) automatically and reliably (#1116).
ess-command now handles sinked consoles correctly.
ess-command no longer changes .Last.value.
As a result, background tasks like completions no longer affect
the last value binding (#1058).
display-buffer is set to pop up frames.
inferior-ess-fix-misaligned-output to t.
xref lookup (M-.).
Function locations are now always detected for package libraries listed
in ess-r-package-library-paths.
examples field. Since roxygen is switching to R markdown, it
becomes useful to evaluate chunks of R outside examples.
ess-set-working-directory no longer changes the active
directory (as defined by the buffer-local variable
default-directory) of the buffer where the command is called.
Instead, the active directory of the inferior buffer is updated to the
new working directory.
'nowait.
With this change you should no longer experience freezes while
evaluating code.
C-c C-e C-r. Reloading now reuses the
same process name and start arguments that were used to start the
process.
with-current-buffer with the return value of the inferior.
:: and ::: operators.
goto-address-mode for url and email highlighting in inferior buffers.
smart-underscore and ess-smart-S-assign-key have been removed.
Users who liked the previous behavior (i.e. underscore inserting “<-”)
should bind ess-insert-assign to the underscore in their Emacs
initialization file. For example, (define-key ess-r-mode-map "_"
#'ess-insert-assign) and (define-key inferior-ess-r-mode-map "_"
#'ess-insert-assign) will activate it in all ESS R buffers.
ess-auto-width-visible controls
visibility.
display-buffer-alist.
Users can now use display-buffer-alist to manage how and where
windows appear. For more information and examples, see See (ess)Controlling
buffer display.
ess-roxy-mode can now be enabled in non-R buffers.
This is primarily intended to support roxygen documentation for cpp
buffers. Preview functionality is not supported outside R buffers.
conf-colon-mode.
ess-style now has effects when set as a file or directory local variable.
ess-default-style is now obsolete, use ess-style instead.
eldoc-echo-area-use-multiline-p
ess-r-rhub-check-package gained new RECOMMENDED.
ess-save-silently.
options(help_type = ``html'') but works with the plain-text
version as well. This only works with options(useFancyQuotes =
TRUE) (the default).
ess-rdired buffers now derive from tabulated-list-mode.
They should look better and be a bit faster overall. The size column
now displays object sizes in bytes.
ess-rdired buffers now auto-update.
The frequency is governed by the new option
ess-rdired-auto-update-interval.
electric-layout-mode is now supported.
This automatically inserts a newline after an opening curly brace in R
buffers. To enable it, customize ess-r-mode-hook.
The following have been made obsolete or removed, see their documentation for more detail:
ess-noweb, ess-swv, and related
functionality like Rnw-mode. Users are encouraged to switch to
one of several other packages that deal with these modes. For example,
polymode https://github.com/polymode/poly-R/,
https://polymode.github.io/, or markdown-mode with edit-indirect
https://jblevins.org/projects/markdown-mode.
auto-complete is obsolete.
The auto-complete package is unmaintained and so ESS support is
now obsolete. Users are encouraged to switch to company-mode
instead.
ess-show-buffer-action,
inferior-ess-same-window, inferior-ess-own-frame, and
inferior-ess-frame-alist. See above about ESS respecting
display-buffer-alist.
ess-tab-always-indent and ess-tab-complete-in-script.
Use the Emacs-wide setting of tab-always-indent instead.
inferior-ess-*-start-file variables.
All modes except Stata did not respect customization of this variable.
In order to load a file on startup, you should put a function on
ess-*-post-run-hook.
Bug Fixes in 18.10.3:
Bug Fixes in 18.10.2:
make install and including
full docs in the tarballs.
Bug Fixes in 18.10.1:
ess-eval-line-visibly-and-step (C-c C-n and ess-eval-region-or-line-visibly-and-step (C-RET)
which behave as the old versions of ess-eval-line-and-step and ess-eval-region-or-line-and-step.
Changes and New Features in 18.10:
package-install does), you should not need to (require
'ess-site) at all, as autoloads should automatically load ESS when it
is needed (e.g. the first time an R buffer is opened). In order to defer
loading your ESS config, you may want to do something like
(with-require-after-load "ess" <ess-config-here>) in your Emacs
init file. Users of the popular use-package Emacs package can now
do (use-package ess :defer t) to take advantage of this behavior.
For more information on this feature, see See (ess)Activating and Loading
ESS.
C-c [letter] bindings have been
removed. This affects C-c h, which was bound to
ess-eval-line-and-step-invisibly in sas-mode-local-map;
C-c f, which was bound to ess-insert-function-outline in
ess-add-MM-keys; and C-c h, which was bound to
ess-handy-commands in Rd-mode-map,
ess-noweb-minor-mode-map, and ess-help-mode-map
ess-eval-line-and-step and ess-eval-region-or-line-and-step
now behave consistently with other evaluation function inside a package.
ess-r-package-use-dir now works with any mode.
This sets the working directory to the root of the current package
including for example C or C++ files within /src).
strip for inferior-ess-replace-long+.
This strips the entire + + sequence.
prog-mode.
In the next release, ESS modes will use define-derived-mode so
that each mode will have (for example) its own hooks and keymaps.
lintr package to use it. Customizable
options include ess-use-flymake, ess-r-flymake-linters,
and ess-r-flymake-lintr-cache.
setwd().
Makevars files open with makefile-mode.
ess-write-to-dribble.
This allows users to disable the dribble (*ESS*) buffer if they wish.
*-program-name variables have been renamed to *-program.
Users who previously customized e.g. inferior-ess-R-program-name
will need to update their customization to
inferior-ess-R-program. These variables are treated as risky
variables.
ess-smart-S-assign was renamed to ess-insert-assign.
It provides similar functionality but for any keybinding, not just ‘_‘.
For instance if you bind it to ‘;‘, repeated invocations cycle through
between assignment and inserting ‘;‘.
C-c C-= is now bound to ess-cycle-assign by default.
See the documentation for details. New user customization option
ess-assign-list controls which assignment operators are cycled.
ess-eval-region for details. This only
works on Emacs 25.1 and newer.
Many improvements to fontification:
ess-R-keywords and ess-R-modifiers.
See the documentation for details.
in is now only fontified when inside a for construct.
This avoids spurious fontification, especially in the output buffer
where ‘in‘ is a common English word.
ess-R-keywords in your Emacs configuration file after loading
ESS (i.e. in the :config section for use-package users).
@param keywords now supports comma-separated parameters.
if () or
stop() are no longer fontified as keyword if not followed by an
opening parenthesis. The same holds for search path modifiers like
library() or require().
%op% operators
and backquoted function definitions.
ess-font-lock-toggle-keyword can be called interactively.
This command asks with completion for a font-lock group to toggle.
This functionality is equivalent to the font-lock menu.
Notable bug fixes:
prettify-symbols-mode no longer breaks indentation.
This is accomplished by having the pretty symbols occupy the same
number of characters as their non-pretty cousins. You may customize
the new variable ess-r-prettify-symbols to control this
behavior.
Obsolete libraries, functions, and variables:
ess-r-args.el library has been obsoleted and will be removed in the next release.
Use eldoc-mode instead, which is on by default.
ess-smart-S-assign,
ess-toggle-S-assign, ess-toggle-S-assign-key,
ess-disable-smart-S-assign.
The variable ess-smart-S-assign-key is now deprecated and will
be removed in the next release. If you would like to continue using
‘_‘ for inserting assign in future releases, please bind
ess-insert-assign in ess-mode-map the normal way.
ess-s-versions-list is obsolete and ignored.
Use ess-s-versions instead. You may pass arguments by starting
the inferior process with the universal argument.
Changes and New Features in 17.11:
(require 'ess-r-mode) or (require 'ess-stata-mode) in your
init file. This is for experienced Emacs users as this requires setting
up autoloads for .R files manually. We will keep maintaining
ess-site for easy loading of all ESS features.
ess-r-package-use-dir. It sets the
working directory of the current process to the current package directory.
ess-r-package-source-roots contains
package sub-directories which are searched recursively during the file
lookup point. Directories in ess-tracebug-search-path are now
also searched recursively.
R/ directory. This way ESS will not attempt to update
function definitions from a package if you are working from e.g. a test
file.
Changes and New Features in 16.10:
%op% operators are
recognized as sexp. This is useful for code navigation, e.g. with
C-M-f and C-M-b.
ess-roxy-fold-examples to non-nil to
automatically fold the examples field when you open a buffer.
ess-roxy-fontify-examples to non-nil to try it out.
ess-r-devtools-ask
bound to C-c C-w C-a. It asks with completion for any devtools
command that takes pkg as argument.
inferior-ess-r-reload-hook
on reloading.
ess-r-package-mode is now activated in non-file buffers as well.
Bug fixes in 16.10:
...= inside list( ).
ess-remote
Changes and New Features in 16.04:
developer functionality has been refactored.
The new user interface consists of a single command
ess-r-set-evaluation-env bound by default to C-c C-t
C-s. Once an evaluation environment has been set with, all subsequent
ESS evaluation will source the code into that environment. By default,
for file within R packages the evaluation environment is set to the
package environment. Set ess-r-package-auto-set-evaluation-env to
nil to disable this.
ess-r-package-mode
This development mode provides features to make package development
easier. Currently, most of the commands are based on the devtools
packages and are accessible with C-c C-w prefix. See the
documentation of ess-r-package-mode function for all available
commands. With C-u prefix each command asks for extra arguments to
the underlying devtools function. This mode is automatically enabled in
all files within R packages and is indicated with [pkg:NAME] in
the mode-line.
options('html_type').
Bug fixes in 16.04:
debug_pipe
The ESS environment is built on the open-source projects of many contributors, dating back to 1989 where Doug Bates and Ed Kademan wrote S-mode to edit S and Splus files in GNU Emacs. Frank Ritter and Mike Meyer added features, creating version 2. Meyer and David Smith made further contributions, creating version 3. For version 4, David Smith provided significant enhancements to allow for powerful process interaction.
John Sall wrote GNU Emacs macros for SAS source code around 1990. Tom Cook added functions to submit jobs, review listing and log files, and produce basic views of a dataset, thus creating a SAS-mode which was distributed in 1994.
In 1994, A.J. Rossini extended S-mode to support XEmacs. Together with extensions written by Martin Maechler, this became version 4.7 and supported S, Splus, and R. In 1995, Rossini extended SAS-mode to work with XEmacs.
In 1997, Rossini merged S-mode and SAS-mode into a single Emacs package for statistical programming; the product of this marriage was called ESS version 5. Richard M. Heiberger designed the inferior mode for interactive SAS and SAS-mode was further integrated into ESS. Thomas Lumley’s Stata mode, written around 1996, was also folded into ESS. More changes were made to support additional statistical languages, particularly XLispStat.
ESS initially worked only with Unix statistics packages that used standard-input and standard-output for both the command-line interface and batch processing. ESS could not communicate with statistical packages that did not use this protocol. This changed in 1998 when Brian Ripley demonstrated use of the Windows Dynamic Data Exchange (DDE) protocol with ESS. Heiberger then used DDE to provide interactive interfaces for Windows versions of Splus. In 1999, Rodney A. Sparapani and Heiberger implemented SAS batch for ESS relying on files, rather than standard-input/standard-output, for Unix, Windows and Mac. In 2001, Sparapani added BUGS batch file processing to ESS for Unix and Windows.
ess-eval-line-and-next-line are by Rod Ball.
ESS is being developed and currently maintained by
If you need to install ESS, read Installing ESS on your system for details on what needs to be done before proceeding to the next chapter. This section describes some of the basics of using Emacs. If you are already familiar with basic Emacs functionality, skip this section. You may also want to use the Emacs tutorial, accessible via C-h t.
In this manual we use the standard notation used by Emacs for describing the keystrokes used to invoke certain commands. C-<chr> means hold the CONTROL key while typing the character <chr>. M-<chr> means hold the META key (usually ALT) down while typing <chr>. If there is no META, EDIT or ALT key, instead press and release the ESC key and then type <chr>.
All ESS commands can be invoked by typing M-x command. Most of the useful commands are bound to keystrokes for ease of use. Also, the most popular commands are also available through the Emacs menubar, and a small subset are provided on the toolbar. Where possible, keybindings are similar to other modes in Emacs to strive for a consistent user interface within Emacs, regardless of the details of which programming language is being edited, or process being run.
Some commands, such as M-x R can accept an optional ‘prefix’ argument. To specify the prefix argument, you would type C-u before giving the command. For example, if you type C-u M-x R, you will be asked for command line options that you wish to invoke the R process with.
Emacs is a ‘self-documenting’ text editor. This applies to ESS in two ways. First, some documentation about each ESS command can be obtained by typing C-h f. For example, if you type C-h f ess-eval-region, documentation for that command will appear in a separate *Help* buffer. Second, C-h m pops up a complete list of keybindings available in each ESS mode and brief description of that mode.
Emacs is a versatile editor written in both C and a dialect of Lisp known as Elisp. ESS is written in Elisp and benefits from the flexible nature of lisp. In particular, many aspects of ESS behaviour can be changed by suitable customization of Lisp variables. This manual mentions some of the most frequent variables. A full list of them however is available by using the Custom facility within Emacs. Type M-x customize-group RET ess RET to get started. Customizing ESS provides details of common user variables you can change to customize ESS to your taste, but it is recommended that you defer this section until you are more familiar with ESS.
ESS supports GNU Emacs versions 25.1 and newer.
ESS is most likely to work with current/recent versions of the following statistical packages: R/S-PLUS, SAS, Stata, OpenBUGS and JAGS.
To build the PDF documentation, you will need a version of TeX Live or texinfo that includes texi2dvi.
There are two main methods used for installing ESS. You may install from a third-party repository or from source code. Once you install it, you must also activate or load ESS in each Emacs session, though installation from a third-party repository likely takes care of that for you. See Activating and Loading ESS for more details.
ESS is packaged by many third party repositories. Many GNU/Linux distributions package it, usually with the name “emacs-ess” or similar.
ESS is also available through Milkypostman’s Emacs Lisp Package Archive (MELPA), a popular repository for Emacs packages. Instructions on how to do so are found on MELPA’s website. MELPA also hosts MELPA-stable with stable ESS builds. You may choose between MELPA with the latest and greatest features (and bugs) or MELPA-stable, which may lag a bit behind but should be more stable.
After installing, users should make sure ESS is activated or loaded in each Emacs session. See Activating and Loading ESS. Depending on install method, this may be taken care of automatically.
Stable versions of ESS are available at the
ESS web page as a .tgz file or .zip
file. ESS releases are GPG-signed, you should check the signature by
downloading the accompanying .sig file and doing:
gpg --verify ess-18.10.tgz.sig
Alternatively, you may download the git repository. ESS is currently
hosted on GitHub: https://github.com/emacs-ess/ESS. git
clone https://github.com/emacs-ess/ESS.git will download it to a new
directory ESS in the current working directory.
We will refer to the location of the ESS source files as /path/to/ESS/ hereafter.
After installing, users should make sure they activate or load ESS in each Emacs session, see Activating and Loading ESS
Optionally, compile Elisp files, build the documentation, and the autoloads:
cd /path/to/ESS/ make
Without this step the documentation, reference card, and autoloads will not be available. Uncompiled ESS will also run slower.
Optionally, you may make ESS available to all users of a machine by
installing it site-wide. To do so, run make install. You might
need administrative privileges:
make install
The files are installed into /usr/share/emacs directory. For this step to
run correctly on macOS, you will need to adjust the PREFIX
path in Makeconf. The necessary code and instructions are
commented in that file.
After installing ESS, you must activate or load it each Emacs session. ESS can be autoloaded, and if you used a third-party repository (such as your Linux distribution or MELPA) to install, you can likely skip this section and proceed directly to Check Installation
Otherwise, you may need to add the path to ESS to load-path with:
(add-to-list 'load-path "/path/to/ESS/lisp")
You then need to decide whether to take advantage of deferred loading
(which will result in a faster Emacs startup time) or require ESS when
Emacs is loaded. To autoload ESS when needed (note that if installed
from source, you must have run make):
(load "ess-autoloads")
To require ESS on startup, you can either put
(require 'ess-site)
or
(require 'ess-r-mode)
In your configuration file, depending on whether you want all ESS features or only R related features.
Restart Emacs and check that ESS was loaded from a correct
location with M-x ess-version.
As well as using ESS to edit your source files for statistical programs, you can use ESS to run these statistical programs. In this chapter, we mostly will refer by example to running R from within Emacs. The Emacs convention is to name such processes running under its control as ‘inferior processes’. Some users find this terminology confusing; you may prefer to think of these as ‘interactive processes.’ Either way, we use the term ‘iESS’ to refer to the Emacs mode used to interact with statistical programs.
To start an inferior R session on GNU/Linux, macOS, or Windows using the Cygwin bash shell, simply type M-x R RET. To start an R session on Windows when you use the MSDOS/powershell shell, simply type M-x S+6-msdos RET. R will then (by default) ask the question
R starting data directory?
Enter the name of the directory you wish to have as the working
directory (that is, the directory you wish to have getwd() return
if using R).
You will then be popped into a buffer named ‘*R*’ which will be used for interacting with the ESS process, and you can start entering commands.
ESS allows you to run more than one iESS process simultaneously in the
same session. Each process has a name and a number; the initial process
(process 1) is simply named ‘R’. If you call M-x R again
without killing the first R process, ESS will start a second R process
with the name ‘R:2’. To have the first buffer named ‘R:1’,
customize the option ess-plain-first-buffername. With a prefix
argument, C-u M-x R allows for the specification of command line
options.
If non-nil, name the first iESS process [R]. Otherwise, name it [R:1].
You can switch to any active ESS process with the command ‘M-x ess-request-a-process’. Just enter the name of the process you require; completion is provided over the names of all running processes. This is a good command to consider binding to a global key.
ESS works with processes on remote computers as easily as with processes on the local machine. The recommended way to access a statistical program on remote computer is to start it with See TRAMP User Manual.
Start an ssh session using TRAMP with ‘C-x C-f /ssh:user@host:
RET’. Tramp should open a dired buffer in your remote home directory.
Now call your favorite ESS process (R, Julia, stata
etc) as you would usually do on local machine: M-x R.
Alternatively you can start your process normally (M-x R). When
asked for starting directory, simply type ‘/ssh:user@host: RET’.
The R process will be started on the remote machine.
To simplify the process even further create a "config" file in your
.ssh/ folder and add an account. For example if you use amazon
EC2, it might look like following:
Host amazon
Hostname ec2-54-215-203-181.us-west-1.compute.amazonaws.com
User ubuntu
IdentityFile ~/.ssh/my_amazon_key.pem
ForwardX11 yes
With this configuration /ssh:amazon: is enough to start a
connection. The ForwardX11 is needed if you want to see the R graphic device
showing on the current machine
TRAMP is the recommended way of starting a remote session. The other
way to start a remote ESS connection is through ess-remote.
Should you or a colleague inadvertently start a statistical process in an ordinary ‘*shell*’ buffer, the ‘ess-remote’ command can be used to convert it to an ESS buffer and allow you to use the ESS commands with it.
If you do not wish ESS to prompt for a starting directory when starting
a new process, set the variable ess-ask-for-ess-directory to
nil. In this case, the starting directory will be set using
one of the following methods:
ess-directory-function stores the name of a
function, the value returned by this function is used. The default for
this variable is nil.
ess-directory stores the name of a
directory (ending in a slash), this value is used. The default for this
variable is nil.
If ess-ask-for-ess-directory has a non-nil value (as it
does by default) then the value determined by the above rules provides
the default when prompting for the starting directory. Incidentally,
ess-directory is an ideal variable to set in
ess-pre-run-hook.
You may also customize ess-post-run-hook and the
dialect-specific ess-*-post-run-hook variables. For example,
you may want to use ess-load-file to load a file when ESS
starts R:
(add-hook 'ess-r-post-run-hook (lambda () (ess-load-file "foo.R")))
If you like to keep a record of your R actions, set the variable
ess-ask-about-transfile to t, and you will be asked for a
filename for the transcript before the ESS process starts.
If non-nil, ask for a file name in which to save the session
transcript.
Enter the name of a file in which to save the transcript at the prompt. If the file doesn’t exist it will be created (for R, you likely want it to end in ‘.Rout’). If the file already exists the transcript will be appended to the file. (Note: if you don’t set this variable but you still want to save the transcript, you can still do it later — see Keeping a record of your R session.)
Once these questions are answered (if they are asked at all) the inferior
process itself is started.
If you need to pass any arguments to this program, they may be specified
in the variable inferior-S_program_name-args. For example, if
inferior-ess-program is "R" then the variable to set is
inferior-R-args.
It is not normally necessary to pass arguments to the iESS program; in
particular do not pass the ‘-e’ option to Splus, since ESS
provides its own command history mechanism.
ESS can set the width option of inferior processes automatically when
the window configuration changes. To do so, set ess-auto-width
to a non-nil value. By default, the change will not be shown in the
inferior buffer. If you want it to be shown, set
ess-auto-width-visible to a non-nil value.
Users can control how buffers are displayed by customizing
display-buffer-alist, See (emacs)Window Choice. This
section provides examples of how to achieve certain setups.
Users coming from RStudio may want the R process to appear at the
bottom left of the frame, help buffers to appear at the bottom right,
and ess-rdired buffers at the top right. To achieve that, you could
this. Note that the order matters; *R Dired* would match
*R if it came before in the alist.
(setq display-buffer-alist
`(("^\\*R Dired"
(display-buffer-reuse-window display-buffer-in-side-window)
(side . right)
(slot . -1)
(window-width . 0.33)
(reusable-frames . nil))
("^\\*R"
(display-buffer-reuse-window display-buffer-at-bottom)
(window-width . 0.5)
(reusable-frames . nil))
("^\\*Help"
(display-buffer-reuse-window display-buffer-in-side-window)
(side . right)
(slot . 1)
(window-width . 0.33)
(reusable-frames . nil))))
Some users prefer working with multiple frames where one frame has the
source code and the other frame has the inferior process. To achieve
this, use this example. For a detailed description of
reusable-frames, see See (elisp)Buffer Display Action Alists.
(setq display-buffer-alist
'(("^\\*R"
(display-buffer-reuse-window display-buffer-pop-up-frame)
(reusable-frames . 0))))
Some users may want the process window to be dedicated to the process so that other buffers are not shown on top of that window. See (elisp)Dedicated Windows. For example, this prevents help buffers from being displayed in the window showing the inferior R process.
(setq display-buffer-alist
'(("^\\*R"
nil
(dedicated . t))))
By default, ESS usually prefers splitting the frame into multiple windows so that you can work with both the R script and the R REPL visible at once. If you want to override this behavior so that the script buffer and the process buffer share the same window, you can do so like this:
(setq display-buffer-alist
'(("^\\*R" . ((display-buffer-same-window) (inhibit-same-window . nil)))
("\\.R$" . ((display-buffer-same-window) (inhibit-same-window . nil)))))
The primary function of the ESS package is to provide an easy-to-use
front end to the R interpreter. This is achieved by running the R
process from within an Emacs buffer, called hereafter inferior
buffer, which has an active inferior-ess-mode. The features of
inferior R mode are similar to those provided by the standard Emacs
shell mode (see Shell Mode in The Gnu Emacs Reference Manual).
Command-line completion of R objects and a number of ‘hot keys’ for
commonly-used R commands are also provided for ease of typing.
Sending a command to the ESS process is as simple as typing it in and pressing the RETURN key:
RET Send the command on the current line to the ESS process.
If you make a typing error before pressing RET all the usual Emacs editing commands are available to correct it (see Basic editing commands in The GNU Emacs Reference Manual). Once the command has been corrected you can press RETURN (even if the cursor is not at the end of the line) to send the corrected command to the ESS process.
Emacs provides some other commands which are useful for fixing mistakes:
backward-kill-word Deletes the previous word (such as an object
name) on the command line.
comint-kill-input Deletes everything from the prompt to point.
Use this to abandon a command you have not yet sent to the ESS process.
comint-bol Move to the beginning of the line, and then skip
forwards past the prompt, if any.
See Shell Mode in The Gnu Emacs Reference Manual, for other commands relevant to entering input.
Most of the time, the cursor spends its time at the bottom of the ESS process buffer, entering commands. However all the input and output from the current (and previous) ESS sessions is stored in the process buffer (we call this the transcript) and often we want to move back up through the buffer, to look at the output from previous commands for example.
Within the process buffer, a paragraph
is defined as the prompt, the command after the prompt, and the output
from the command. Thus M-{ and M-} move you backwards and
forwards, respectively, through commands in the transcript. A
particularly useful command is M-h (mark-paragraph) which
will allow you to mark a command and its entire output (for deletion,
perhaps). For more information about paragraph commands,
see Paragraphs in The GNU Emacs
Reference Manual.
If an ESS process finishes and you restart it in the same process buffer, the output from the new ESS process appears after the output from the first ESS process separated by a form-feed (‘^L’) character. Thus pages in the ESS process buffer correspond to ESS sessions. Thus, for example, you may use C-x [ and C-x ] to move backward and forwards through ESS sessions in a single ESS process buffer. For more information about page commands, see Pages in The GNU Emacs Reference Manual.
Viewing the output of the command you have just entered is a common occurrence and ESS provides a number of facilities for doing this. Whenever a command produces output, it is possible that the window will scroll, leaving the next prompt near the middle of the window. The first part of the command output may have scrolled off the top of the window, even though the entire output would fit in the window if the prompt were near the bottom of the window. If this happens, you can use the following comint commands:
comint-show-maximum-output to move to the end
of the buffer, and place cursor on bottom line of window to make more of
the last output visible. To make this happen automatically for all
inputs, set the variable comint-scroll-to-bottom-on-input to
t or 'this. If the first part of the output is still not
visible, use
C-c C-r (comint-show-output),
which moves cursor to the previous command line and places it at the
top of the window.
Finally, if you want to discard the last command output altogether, use
C-c C-o (comint-delete-output),
which deletes everything from the last command to the current prompt.
Use this command judiciously to keep your transcript to a more
manageable size.
If you want to view the output from more historic commands than the previous command, commands are also provided to move backwards and forwards through previously entered commands in the process buffer:
comint-previous-prompt Moves point to the preceding prompt in the
process buffer.
comint-next-prompt Moves point to the next prompt in the process
buffer.
Note that these two commands are analogous to C-p and C-n but apply to command lines rather than text lines. And just like C-p and C-n, passing a prefix argument to these commands means to move to the ARG’th next (or previous) command. (These commands are also discussed in Shell History Copying in The GNU Emacs Reference Manual.)
There are also two similar commands (not bound to any keys by default) which move to preceding or succeeding commands, but which first prompt for a regular expression (see Syntax of Regular Expression in The GNU Emacs Reference Manual), and then moves to the next (previous) command matching the pattern.
comint-backward-matching-input regexp arg ¶comint-forward-matching-input regexp argSearch backward (forward) through the transcript buffer for the arg’th previous (next) command matching regexp. arg is the prefix argument; regexp is prompted for in the minibuffer.
When moving through the transcript, you may wish to re-execute some of the commands you find there. ESS provides commands to do this; these commands may be used whenever the cursor is within a command line in the transcript (if the cursor is within some command output, an error is signaled). Note all commands involve the RETURN key.
inferior-ess-send-input See Entering commands and fixing mistakes.
comint-copy-old-input Copy the command under the cursor to the
current command line, but don’t execute it. Leaves the cursor on the
command line so that the copied command may be edited.
When the cursor is not after the current prompt, the RETURN key has a slightly different behavior than usual. Pressing RET on any line containing a command that you entered (i.e. a line beginning with a prompt) sends that command to the ESS process once again. If you wish to edit the command before executing it, use C-c RET instead; it copies the command to the current prompt but does not execute it, allowing you to edit it before (re)submitting it.
These commands work even if the current line is a continuation line (i.e. the prompt is ‘+’ instead of ‘>’) — in this case all the lines that form the multi-line command are concatenated together and the resulting command is sent to the ESS process (currently this is the only way to resubmit a multi-line command to the ESS process in one go). If the current line does not begin with a prompt, an error is signalled. This feature, coupled with the command-based motion commands described above, could be used as a primitive history mechanism. ESS provides a more sophisticated mechanism, however, which is described in Command History.
To keep a record of your R session in a disk file, use the Emacs
command C-x C-w (write-file) to attach a file to the ESS
process buffer. The name of the process buffer will (probably) change
to the name of the file, but this is not a problem. You can still use
R as usual; just remember to save the file before you quit Emacs
with C-x C-s. You can make ESS prompt you for a filename in which
to save the transcript every time you start R by setting the
variable
ess-ask-about-transfile to t; See Changing the startup actions.
For R files, naming transcript files ‘*.Rout’ puts them in a
special mode (ESS transcript mode — see Manipulating saved transcript files) for
editing transcript files which is automatically selected for files with
this suffix.
R transcripts can get very large, so some judicious editing is appropriate if you are saving it in a file. Use C-c C-o whenever a command produces excessively long output (printing large arrays, for example). Delete erroneous commands (and the resulting error messages or other output) by moving to the command (or its output) and typing M-h C-w. Also, remember that C-c C-x (and other hot keys) may be used for commands whose output you do not wish to appear in the transcript. These suggestions are appropriate even if you are not saving your transcript to disk, since the larger the transcript, the more memory your Emacs process will use on the host machine.
You can use ess-transcript-clean-region to strip output from a
transcript, leaving only source code suitable for inclusion in files
source()-able from R. see Transcript Mode
ESS provides easy-to-use facilities for re-executing or editing previous
commands. An input history of the last few commands is maintained (by
default the last 500 commands are stored, although this can be changed by
setting the variable comint-input-ring-size in
inferior-ess-mode-hook.) The simplest history commands simply
select the next and previous commands in the input history:
comint-previous-input Select the previous command in the input
history.
comint-next-input Select the next command in the input history.
For example, pressing M-p once will re-enter the last command into the process buffer after the prompt but does not send it to the ESS process, thus allowing editing or correction of the command before the ESS process sees it. Once corrections have been made, press RET to send the edited command to the ESS process.
If you want to select a particular command from the history by matching it against a regular expression (see Syntax of Regular Expression in The GNU Emacs Reference Manual), to search for a particular variable name for example, these commands are also available:
comint-history-isearch-backward-regexp Prompt for a regular
expression, and search backwards through the input history for a command
matching the expression.
A common type of search is to find the last command that began with a particular sequence of characters; the following two commands provide an easy way to do this:
comint-previous-matching-input-from-input Select the previous
command in the history which matches the string typed so far.
comint-next-matching-input-from-input Select the next command in
the history which matches the string typed so far.
Instead of prompting for a regular expression to match against, as they
instead select commands starting with those characters already entered.
For instance, if you wanted to re-execute the last attach()
command, you may only need to type att and then C-c M-r and
RET.
See Shell History Ring in The GNU Emacs Reference
Manual, for a more detailed discussion of the history mechanism,
and do experiment with the In/Out menu to explore the possibilities.
Many ESS users like to have even easier access to these, and recommend adding something like
(eval-after-load "comint"
'(progn
(define-key comint-mode-map [up]
'comint-previous-matching-input-from-input)
(define-key comint-mode-map [down]
'comint-next-matching-input-from-input)
;; also recommended for ESS use --
(setq comint-move-point-for-output 'others)
;; somewhat extreme, almost disabling writing in *R*, *shell* buffers above prompt:
(setq comint-scroll-to-bottom-on-input 'this)
))
to your Emacs configuration file, where the last two settings are
typically desirable for the situation where you work with a script (for
example, code.R) and send code chunks to the process buffer
(e.g. *R*). Note however that these settings influence all
comint-using Emacs modes, not just the ESS ones, and for that
reason, these customization cannot be part of ESS itself.
The ess-history-file variable, which is t by default,
together with ess-history-directory, governs if and where the
command history is saved and restored between sessions.
By default, ess-history-directory is nil, and the command
history will be stored (as a file) in the same directory as the iESS process.
ESS users may work exclusively with script files rather than in a iESS session, and may not want to save any history files. To do so:
(setq ess-history-file nil)
or if you prefer only one global command history file:
(setq ess-history-directory "~/.R/")
in your Emacs configuration file.
Instead of searching through the command history using the command
described in the previous section, you can alternatively refer to a
historical command directly using a notation very similar to that used
in csh. History references are introduced by a ‘!’ or
‘^’ character and have meanings as follows:
The immediately previous command
The Nth previous command
The last command beginning with the string ‘text’
The last command containing the string ‘text’
In addition, you may follow the reference with a word designator
to select particular words of the input. A word is defined as a
sequence of characters separated by whitespace. (You can modify this
definition by setting the value of comint-delimiter-argument-list
to a list of characters that are allowed to separate words and
themselves form words.) Words are numbered beginning with zero. The
word designator usually begins with a ‘:’ (colon) character;
however it may be omitted if the word reference begins with a ‘^’,
‘$’, ‘*’ or ‘-’. If the word is to be selected from the
previous command, the second ‘!’ character can be omitted from the
event specification. For instance, ‘!!:1’ and ‘!:1’ both
refer to the first word of the previous command, while ‘!!$’ and
‘!$’ both refer to the last word in the previous command. The
format of word designators is as follows:
The zeroth word (i.e. the first one on the command line)
The nth word, where n is a number
The first word (i.e. the second one on the command line)
The last word
A range of words; ‘-y’ abbreviates ‘0-y’
All the words except the zeroth word, or nothing if the command had just one word (the zeroth)
Abbreviates x-$
Like ‘x*’, but omitting the last word
In addition, you may surround the entire reference except for the first ‘!’ by braces to allow it to be followed by other (non-whitespace) characters (which will be appended to the expanded reference).
Finally, ESS also provides quick substitution; a reference like ‘^old^new^’ means “the last command, but with the first occurrence of the string ‘old’ replaced with the string ‘new’” (the last ‘^’ is optional). Similarly, ‘^old^’ means “the last command, with the first occurrence of the string ‘old’ deleted” (again, the last ‘^’ is optional).
To convert a history reference as described above to an input suitable for R, you need to expand the history reference, using the TAB key. For this to work, the cursor must be preceded by a space (otherwise it would try to complete an object name) and not be within a string (otherwise it would try to complete a filename). So to expand the history reference, type SPC TAB. This will convert the history reference into an R command from the history, which you can then edit or press RET to execute.
For example, to execute the last command that referenced the variable
data, type !?data SPC TAB RET.
ESS provides a number of commands for executing the commonly used
functions. These commands below are basically information-gaining
commands (such as objects() or search()) which tend to
clutter up your transcript and for this reason some of the hot keys
display their output in a temporary buffer instead of the process buffer
by default. This behavior is controlled by the following option:
If non-nil, means that these commands will produce their output
in the process buffer instead.
In any case, passing a prefix argument to the commands (with C-u)
will reverse the meaning of ess-execute-in-process-buffer for
that command. In other words, the output will be displayed in the
process buffer if it usually goes to a temporary buffer, and vice-versa.
These are the hot keys that behave in this way:
C-c C-x Sends the objects() command to the ESS process. A
prefix argument specifies the position on the search list (use a
negative argument to toggle ess-execute-in-process-buffer as
well). This is a quick way to see what objects are in your working
directory. A prefix argument of 2 or more means get objects for that
position. A negative prefix argument posn gets the objects for
that position, as well as toggling ess-execute-in-process-buffer.
C-c C-s Sends the inferior-ess-search-list-command command
to the ess-language process; search() in R. Prefix
invert toggles ess-execute-in-process-buffer.
ess-execute may seem pointless when you could just type the
command in anyway, but it proves useful for ‘spot’ calculations which
would otherwise clutter your transcript, or for evaluating an expression
while partway through entering a command. You can also use this command
to generate new hot keys using the Emacs keyboard macro facilities;
see Keyboard Macros in The GNU
Emacs Reference Manual.
The following hot keys do not use ess-execute-in-process-buffer
to decide where to display the output — they either always display in
the process buffer or in a separate buffer, as indicated:
C-c M-l Prompts for a file (filename) to load into the ESS
process using source(). If there is an error during loading, you
can jump to the error in the file with the following function.
Visits next next-error message and
corresponding source code. If all the error messages parsed so far have
been processed already, the message buffer is checked for new ones. A
prefix arg specifies how many error messages to move; negative
means move back to previous error messages. Just C-u as a prefix
means reparse the error message buffer and start at the first error.
The reset argument specifies restarting from the beginning.
See Detecting errors in source files, for more details.
C-c C-v Pops up a help buffer for an R object or
function. If command is supplied, it is used instead of
inferior-ess-help-command. See Reading help files for more details.
C-c C-q Issue an exiting command to the inferior process,
additionally also running ess-cleanup for disposing of any
temporary buffers (such as help buffers and edit buffers) that may have
been created. Use this command when you have finished your R
session instead of simply quitting at the inferior process prompt,
otherwise you will need to issue the command ess-cleanup
explicitly to make sure that all the files that need to be saved have
been saved, and that all the temporary buffers have been killed.
For the R languages (R, S, S-Plus) ESS sets an option in the current process that programs in the language can check to determine the environment in which they are currently running.
ESS sets options(STERM="iESS") for R language processes running
in an inferior iESS[S] or iESS[R] buffer.
ESS sets options(STERM="ddeESS") for independent S-Plus for
Windows processes running in the GUI and communicating with ESS via the
DDE (Microsoft Dynamic Data Exchange) protocol through a
ddeESS[S] buffer.
Other values of options()$STERM that we recommend are:
length: Fixed length xterm or telnet window.
scrollable: Unlimited length xterm or telnet window.
server: S-Plus Stat Server.
BATCH: BATCH.
Rgui: R GUI.
Commands: S-Plus GUI without DDE interface to ESS.
Additional values may be recommended in the future as new interaction
protocols are created. Unlike the values iESS and ddeESS,
ESS can’t set these other values since the R language program is not
under the control of ESS.
When starting R or S under Unix, ESS sets
options(editor="emacsclient"). Under Microsoft Windows, it will
use gnuclient.exe rather than emacsclient, but the same principal
applies. Within your R session, if you have a function
called iterator, typing fix(iterator), will show that
function in a temporary Emacs buffer. You can then correct the
function. When you kill the buffer, the definition of the function is
updated. Using edit() rather than fix() means that the
function is not updated. Finally, the R function page(x) will
also show a text representation of the object x in a temporary
Emacs buffer.
The following commands are also available in the process buffer:
C-c C-c Sends a Control-C signal to the iESS process. This has the effect of aborting the current command.
C-c C-z When in process buffer, return to the most recent script buffer. When in a script buffer pop to the associated process buffer. Consecutive presses of C-z switch between the script and process buffers.
If toggle-eob is given, the value of
ess-switch-to-end-of-proc-buffer is toggled.
If non-nil, ess-switch-to-inferior-or-script-buffer goes
to end of process buffer.
Other commands available in iESS modes are discussed in Shell Mode in The Gnu Emacs Reference Manual.
Other commands are also available for evaluating portions of code in the
R process. These commands cause the selected code to be evaluated
directly by the ESS process as if you had typed them in at the
command line; the source() function is not used. You may choose
whether both the commands and their output appear in the process buffer
(as if you had typed in the commands yourself) or if the output alone is
echoed. The behavior is controlled by the variable:
Non-nil means ess-eval-* commands display commands and
output in the process buffer. A value of t blocks Emacs while R
is busy. A value of nowait does not block Emacs but printing
may be slightly off.
Passing a prefix (C-u) (in Elisp terms, the argument VIS) to any of the following commands,
however, reverses the meaning of ess-eval-visibly for that
command only — for example C-u C-c C-j evaluates the current
line without showing the input in the iESS buffer. The default value of
ess-eval-visibly (t) means that ESS calls block Emacs
until they finish. This may be undesirable, especially if commands take
long to finish. Users who want input to be displayed and Emacs not to be
blocked can set ess-eval-visibly to 'nowait. This sends
the input to the iESS buffer but does not wait for the process to
finish, ensuring Emacs is not blocked.
Primary commands for evaluating code are:
C-<RET> Sends the highlighted region or current line and step to next line of code.
C-M-x Sends the current selected region or function or paragraph.
C-c C-c Like ess-eval-region-or-function-or-paragraph but
steps to next line of code.
Other, not so often used, evaluation commands are:
C-c C-j Sends the current line to the ESS process.
C-c M-j Like ess-eval-line but additionally switches point
to the ESS process.
C-c C-f Sends the function containing point to the ESS process.
C-c M-f Like ess-eval-function but additionally switches
point to the ESS process.
C-c C-r Sends the current region to the ESS process.
C-c M-r Like ess-eval-region but additionally switches
point to the ESS process.
C-c C-b Sends the current buffer to the ESS process.
C-c M-b Like ess-eval-buffer but additionally switches
point to the ESS process.
All the above ess-eval-* commands are useful for evaluating small
amounts of code and observing the results in the process buffer for
debugging purposes, or for generating transcripts from source files.
When editing R functions, it is generally preferable to use C-c
C-l to update the function’s value. In particular,
ess-eval-buffer is now largely obsolete.
A useful way to work is to divide the frame into two windows; one
containing the source code and the other containing the process buffer.
If you wish to make the process buffer scroll automatically when the
output reaches the bottom of the window, you will need to set the
variable comint-move-point-for-output to 'others or
t.
Inferior R mode records the transcript (the list of all commands executed, and their output) in the process buffer, which can be saved as a transcript file, which should normally have the suffix .Rout. The most obvious use for a transcript file is as a static record of the actions you have performed in a particular R session. Sometimes, however, you may wish to re-execute commands recorded in the transcript file by submitting them to a running ESS process. This is what Transcript Mode is for.
If you load file a with the suffix .Rout into Emacs, it is placed in R Transcript Mode. Transcript Mode is similar to inferior R mode (see Interacting with the ESS process): paragraphs are defined as a command and its output, and you can move though commands either with the paragraph commands or with C-c C-p and C-c C-n.
Three commands are provided to re-submit command lines from the transcript file to a running ESS process. They are:
M-RET Sends the current command line to the ESS process, and execute it.
C-c RET Copy the current command to the ESS process, and switch to it (ready to edit the copied command).
RET Sends the current command to the ESS process, and move to the next command line. This command is useful for submitting a series of commands.
Note that the first two commands are similar to those on the same keys in inferior R Mode. In all three cases, the commands should be executed when the cursor is on a command line in the transcript; the prompt is automatically removed before the command is submitted.
Yet another use for transcript files is to extract the command lines for inclusion in an R source file or function. Transcript mode provides one command which does just this:
C-c C-w Strip the transcript in the region (given by beg and end), leaving only commands. Deletes any lines not beginning with a prompt, and then removes the prompt from those lines that remain. Prefix argument even-if-read-only means to clean even if the buffer is read-only. Don’t forget to remove any erroneous commands first!
The remaining command lines may then be copied to a source file or edit buffer for inclusion in a function definition, or may be evaluated directly (see Sending code to the ESS process) using the code evaluation commands from R mode, also available in R Transcript Mode.
ESS provides facilities for editing R objects within your Emacs session. Most editing is performed on R functions, although in theory you may edit datasets as well. Edit buffers are always associated with files, although you may choose to make these files temporary if you wish. Alternatively, you may make use of a simple yet powerful mechanism for maintaining backups of text representations of R functions. Error-checking is performed when R code is loaded into the ESS process.
To edit an object, type
C-c C-e C-d Edit an object in its own edit buffer.
from within the iESS process buffer (*R*). You will then be
prompted for an object to edit: you may either type in the name of an
existing object (for which completion is available using the TAB
key),
or you may enter the name of a new object.
A buffer will be created containing the text representation of the
requested object or, if you entered the name of a non-existent object at
the prompt and the variable ess-function-template
is non-nil, you will be presented with a template defined by that
variable, which defaults to a skeleton function construct.
You may then edit the function as required. The edit buffer generated
by ess-dump-object-into-edit-buffer is placed in the ESS
major mode which provides a number of commands to facilitate editing
R source code. Commands are provided to intelligently indent R
code, evaluate portions of R code and to move around R code
constructs.
Note: when you dump a file with C-c C-e C-d, ESS first checks
to see whether there already exists an edit buffer containing that
object and, if so, pops you directly to that buffer. If not, ESS next
checks whether there is a file in the appropriate place with the
appropriate name (see Maintaining R source files) and if so, reads in that file.
You can use this facility to return to an object you were editing in a
previous session (and which possibly was never loaded to the R
session). Finally, if both these tests fail, the ESS process is
consulted and a dump() command issued.
If you want to force ESS to ask the ESS process for the object’s
definition (say, to reformat an unmodified buffer or to revert back to
R’s idea of the object’s definition) pass a prefix argument to
ess-dump-object-into-edit-buffer by typing C-u C-c C-e C-d.
The best way to get information — particularly function definitions
— into R is to load them in as source file, using R’s
source function. You have already seen how to create source
files using C-c C-e C-d; ESS provides a complementary command for
loading source files (even files not created with ESS!) into the ESS
process, namely ess-load-file (C-c M-l).
see Hot keys for common commands.
After typing C-c M-l you will prompt for the name of the file to load into R; usually this is the current buffer’s file which is the default value (selected by simply pressing RET at the prompt). You will be asked to save the buffer first if it has been modified (this happens automatically if the buffer was generated with C-c C-e C-d). The file will then be loaded, and if it loads successfully you will be returned to the ESS process.
If any errors occur when loading a file with C-c C-l, ESS will
inform you of this fact. In this case, you can jump directly to the
line in the source file which caused the error by typing C-c `
(ess-parse-errors).
You will be returned to the offending file (loading it into a buffer if
necessary) with point at the line S reported as containing the
error. You may then correct the error, and reload the file. Note that
none of the commands in an R source file will take effect if any
part of the file contains errors.
Sometimes the error is not caused by a syntax error (loading a
non-existent file for example). In this case typing C-c ` will
simply display a buffer containing S’s error message. You can force
this behavior (and avoid jumping to the file when there is a
syntax error) by passing a prefix argument to ess-parse-errors
with C-u C-c `.
ESS provides a sophisticated mechanism for indenting R source
code. Compound statements (delimited by
‘{’ and ‘}’) are indented relative to their enclosing block.
In addition, the braces have been electrified to automatically indent to
the correct position when inserted, and optionally insert a newline at
the appropriate place as well. Lines which continue an incomplete
expression are indented relative to the first line of the expression.
Function definitions, if statements, calls to expression()
and loop constructs are all recognized and indented appropriately. User
variables are provided to control the amount of indentation in each
case, and there are also a number of predefined indentation styles to
choose from.
Comments are also handled specially by ESS, using an idea borrowed from
the Emacs-Lisp indentation style. By default,
comments beginning with ‘###’
are aligned to the beginning of the line. Comments beginning with
‘##’ are aligned to the current level of indentation for the block
containing the comment. Finally, comments beginning with ‘#’ are
aligned to a column on the right (the 40th column by default, but this
value is controlled by the variable comment-column,)
or just after the expression on the line containing the comment if it
extends beyond the indentation column. You turn off the default
behavior by adding the line
(setq ess-indent-with-fancy-comments nil) to your .emacs file.
ESS also supports roxygen2 entries which is R documentation maintained in the source code as comments See Editing roxygen2 documentation.
The indentation commands provided by ESS are:
TAB Indents the current line as R code.
Try to indent first, and if code is already properly indented, complete
instead. In ess-mode, only tries completion if
tab-always-indent is 'complete. See also
ess-first-tab-never-complete.
If non-nil, TAB never tries to complete in ess-mode. The
default 'symbol does not try to complete if the next char is a valid
symbol constituent. There are more options, see the help (C-h v).
TAB Indents each line in the R (compound) expression which follows point. Very useful for beautifying your R code.
{ } The braces automatically indent to the correct position when typed.
The following Emacs command are also helpful:
newline-and-indent Insert a newline, and indent the next line.
(Note that most keyboards nowadays do not have a LINEFEED key, but
RET and C-j are equivalent.)
indent-dwim Call the comment command you want (Do What I Mean).
The combined value of twelve variables (4 of three groups ess-indent-*,
ess-offset-* and ess-align-*) that control indentation are
collectively termed a style. ESS provides several styles covering the
common styles of indentation: DEFAULT, OWN, GNU,
BSD, K&R, C++, RRR, RRR+,
Rstudio, Rstudio-, and CLB. The
variable ess-style-alist lists the value of each indentation
variable per style (and its documentation contains more).
C-c C-e C-s (or C-c C-e s) sets the formatting style in this
buffer to be one of the predefined styles, see above. The DEFAULT
style uses the default values for the indenting variables; The OWN
style allows you to use your own private values of the indentation
variable, see below.
The style to use. See the variable ess-style-alist for how
these groups (DEFAULT, OWN, GNU, RRR, …) map onto different
settings for variables. You can set this in your Emacs configuration
file:
(setq ess-style 'C++)
Predefined formatting styles for ESS code. Values for all groups,
except OWN, are fixed. To change the value of variables in the OWN
group, customize the variable ess-own-style-list. The default
style in use is controlled by ess-style.
The styles DEFAULT and OWN are initially identical. If
you wish to edit some of the default values, set
ess-style to 'OWN and change
ess-own-style-list.
See Customizing ESS, for convenient ways to set both these variables.
If you prefer not to use the customization facility, you can change individual indentation variables within a hook, for example:
(defun myindent-ess-hook () (setq ess-indent-level 4)) (add-hook 'ess-mode-hook 'myindent-ess-hook)
In the rare case that you’d like to add an entire new indentation style
of your own, copy the definition of ess-own-style-list to a new
variable and ensure that the last line of the :set declaration
calls ess-add-style with a unique name for your style
(e.g. 'MINE). Finally, add (setq ess-style 'MINE)
to use your new style.
A number of commands are provided to move across function definitions in the edit buffer:
ESC C-a aka C-M-a If inside a function go to the beginning of it, otherwise go to the beginning of paragraph.
ESC C-e aka C-M-e Move point to the end of the function containing point.
ESC C-h aka C-M-h Place point at the beginning of the R function containing point, and mark at the end.
Don’t forget the usual Emacs commands for moving over balanced expressions and parentheses: See Lists and Sexps in The GNU Emacs Reference Manual.
Completion is provided in the edit buffer in a similar fashion to the process buffer: TAB first indents, and if there is nothing to indent, completes the object or file name; M-? lists file completions. See See Completion, for more.
Finally, C-c C-z (ess-switch-to-inferior-or-script-buffer)
returns you to the iESS process buffer, if done from a script
buffer, placing point at the end of the buffer. If this is done from
the iESS process buffer, point is taken to the script buffer.
In addition many commands available in the process buffer are also available in the script buffer. You can still read help files with C-c C-v, edit another function with C-c C-e C-d and of course C-c C-l can be used to load a source file into R.
Every edit buffer in ESS is associated with a dump file on disk.
Dump files are created whenever you type C-c C-e C-d
(ess-dump-object-into-edit-buffer), and may either be deleted
after use, or kept as a backup file or as a means of keeping several
versions of an R function.
If non-nil, dump files created with C-c C-e C-d are deleted
immediately after they are created by the ess-process.
Since immediately after R dumps an object’s definition to a disk
file the source code on disk corresponds exactly to R’s idea of the
object’s definition, the disk file isn’t needed; deleting it now has the
advantage that if you don’t modify the file (say, because you
just wanted to look at the definition of one of the standard R
functions) the source dump file won’t be left around when you kill the
buffer. Note that this variable only applies to files generated with
R’s dump function; it doesn’t apply to source files which already
exist. The default value is t.
Variable controlling whether to delete dump files after a successful
load. If ‘nil’: always delete. If ‘ask’, confirm to delete.
If ‘check’, confirm to delete, except for files created with
ess-dump-object-into-edit-buffer. Anything else, never delete.
This variable only affects the behaviour of ess-load-file. Dump
files are never deleted if an error occurs during the load.
After an object has been successfully (without error) loaded back into R with C-c C-l, the disk file again corresponds exactly (well, almost — see below) to R’s record of the object’s definition, and so some people prefer to delete the disk file rather than unnecessarily use up space. This option allows you to do just that.
If the value of ess-keep-dump-files is t, dump files are
never deleted after they are loaded. Thus you can maintain a complete
text record of the functions you have edited within ESS. Backup files
are kept as usual, and so by using the Emacs numbered backup facility
— see Single or Numbered
Backups in The Gnu Emacs Reference Manual, you can keep a historic
record of function definitions. Another possibility is to maintain the
files with a version-control system such as git See Version Control in The Gnu Emacs Reference
Manual. As long as a dump file exists in the appropriate place for a
particular object, editing that object with C-c C-e C-d finds that
file for editing (unless a prefix argument is given) — the ESS process
is not consulted. Thus you can keep comments outside the
function definition as a means of documentation that does not clutter
the R object itself. Another useful feature is that you may format the
code in any fashion you please without R re-indenting the code every
time you edit it. These features are particularly useful for
project-based work.
If the value of ess-keep-dump-files is nil, the dump file is always
silently deleted after a successful load with C-c C-l. While this
is useful for files that were created with C-c C-e C-d it also applies
to any other file you load (say, a source file of function
definitions), and so can be dangerous to use unless you are careful.
Note that since ess-keep-dump-files is buffer-local, you can make
sure particular files are not deleted by setting it to t in the
Local Variables section of the file See Local Variables in Files in The Gnu Emacs
Reference Manual.
A safer option is to set ess-keep-dump-files to ask; this
means that ESS will always ask for confirmation before deleting the
file. Since this can get annoying if you always want to delete dump
files created with C-c C-e C-d, but not any other files, setting
ess-keep-dump-files to check (the default value) will
silently delete dump files created with C-c C-e C-d in the current
Emacs session, but query for any other file. Note that in any case you
will only be asked for confirmation once per file, and your answer is
remembered for the rest of the Emacs session.
Note that in all cases, if an error (such as a syntax error) is detected while loading the file with C-c C-l, the dump file is never deleted. This is so that you can edit the file in a new Emacs session if you happen to quit Emacs before correcting the error.
Dump buffers are always autosaved, regardless of the value of
ess-keep-dump-files.
Every dump file should be given a unique file name, usually the dumped object name with some additions.
Template for filenames of dumped objects. %s is replaced by the
object name.
By default, dump file names are the user name, followed by ‘.’ and
the object and ending with ‘.R’. Thus if user joe dumps the
object myfun the dump file will have name joe.myfun.R.
The username part is included to avoid clashes when dumping into a
publicly-writable directory, such as /tmp; you may wish to remove
this part if you are dumping into a directory owned by you.
You may also specify the directory in which dump files are written:
Directory name (ending in a slash) where R dump files are to be written.
By default, dump files are always written to /tmp, which is fine
when ess-keep-dump-files is nil. If you are keeping dump
files, then you will probably want to keep them somewhere in your home
directory, say ~/R-source. This could be achieved by including
the following line in your Emacs configuration file:
(setq ess-source-directory (expand-file-name "~/R-source/"))
If you would prefer to keep your dump files in separate directories
depending on the value of some variable, ESS provides a facility for
this also. By setting ess-source-directory to a lambda
expression which evaluates to a directory name, you have a great deal of
flexibility in selecting the directory for a particular source file to
appear in. The lambda expression is evaluated with the process buffer
as the current buffer and so you can use the variables local to that
buffer to make your choice. For example, the following expression
causes source files to be saved in the subdirectory Src of the
directory the ESS process was run in.
(setq ess-source-directory
(lambda ()
(concat ess-directory "Src/")))
(ess-directory is a buffer-local variable in process buffers
which records the directory the ESS process was run from.) This is
useful if you keep your dump files and you often edit objects with the
same name in different ESS processes. Alternatively, if you often
change your R working directory during an R session, you may
like to keep dump files in some subdirectory of the directory pointed to
by the first element of the current search list. This way you can edit
objects of the same name in different directories during the one R
session:
(setq ess-source-directory
(lambda ()
(file-name-as-directory
(expand-file-name (concat
(car ess-search-list)
"/.Src")))))
If the directory generated by the lambda function does not exist but can be created, you will be asked whether you wish to create the directory. If you choose not to, or the directory cannot be created, you will not be able to edit functions.
ESS provides an easy-to-use facility for reading R help files from
within Emacs. From within the ESS process buffer or any ESS edit
buffer, typing C-c C-v (ess-display-help-on-object)
will prompt you for the name of an object for which you would like
documentation. Completion is provided over all objects which have help
files.
If the requested object has documentation, you will be popped into a
buffer (named *help(obj-name)*) containing the help file.
This buffer is placed in a special ESS help mode which disables the
usual editing commands but which provides a number of keys for paging
through the help file.
Help commands:
ess-describe-help-mode Pops up a help buffer with a list of the
commands available in R help mode.
ess-display-help-on-object Pop up a help buffer for a different
object.
scroll-down-command Move one screen backwards through the help file.
scroll-up-command Move one screen forwards through the help file.
end/beginning-of-buffer Move to the end or beginning of the help file,
respectively.
Section-based motion commands:
ess-skip-to-previous-section and ess-skip-to-next-section
Move to the next and previous section header in the help file,
respectively. A section header consists of a number of capitalized
words, followed by a colon.
In addition, the s key followed by one of the following letters will jump to a particular section in the help file. Note that the exact headings available and capitalization scheme may vary across languages.
You may use s ? to get the current list of active key bindings.
Arguments:
Background:
Bugs:
Description:
Details:
Examples:
Note:
Optional Arguments:
Required Arguments:
References:
See Also:
Side Effects:
Usage:
Value:
Pops up a help buffer with a list of the defined section motion keys.
Evaluation:
ess-eval-line-and-step Evaluates the current line in the ESS
process, and moves to the next line. Useful for running examples in
help files.
ess-eval-region Send the contents of the current region to the
ESS process. Useful for running examples in help files.
Quit commands:
ess-help-quit Return to previously selected buffer, and bury the
help buffer.
kill-buffer Return to previously selected buffer, and kills the
help buffer.
ess-kill-buffer-and-go Return to the ESS process, killing this
help buffer.
Miscellaneous:
ess-display-index Prompt for a package and display it’s help
index.
ess-display-vignettes Display all available vignettes.
ess-display-help-in-browser Display current help page with the
web browser.
isearch-forward Same as C-s.
In addition, all of the ESS commands available in the edit buffers are also available in R help mode (see Creating or modifying R objects). Of course, the usual (non-editing) Emacs commands are available, and for convenience the digits and - act as prefix arguments.
If a help buffer already exists for an object for which help is
requested, that buffer is popped to immediately; the ESS process is
not consulted at all. If the contents of the help file have changed,
you either need to kill the help buffer first, or pass a prefix argument
(with C-u) to ess-display-help-on-object.
Help buffers are marked as temporary buffers in ESS, and are deleted
when ess-quit or ess-cleanup are called.
Help buffers normally appear in another window within the current
frame. If you wish help buffers to appear in their own frame (either
one per help buffer, or one for all help buffers), you can customize the
variable ess-help-own-frame.
Help buffers are displayed by calling the function
ess-display-help. You can customize where these buffers are
displayed by adding an entry in display-buffer-alist; for
examples, see See (ess)Controlling buffer display. Or by
customizing the options ess-help-own-frame,
ess-help-frame-alist, and ess-display-buffer-reuse-frame.
The TAB key is for completion. The value of the variable
ess-first-tab-never-complete controls when completion is allowed to
occur. In ess-mode TAB first tries to indent, and if there is
nothing to indent, complete the object name instead.
comint-dynamic-complete Complete the R object name or
filename before point.
When the cursor is just after a partially-completed object name,
pressing TAB provides completion in a similar fashion to the rest of Emacs. ESS maintains a list of all objects known to R
at any given time, which basically consists of all objects (functions
and datasets) in every attached directory listed by the search()
command
along with the component objects of attached data frames
For example, consider three functions binomplot(),
binom.test() and binomial(). Typing bin TAB will
insert the characters ‘om’, completing the longest prefix
(‘binom’) which distinguishes these three commands. Pressing
TAB once more provides a list of the three commands which have
this prefix, allowing you to add more characters (say, ‘.’) which
specify the function you desire. After entering more characters
pressing TAB yet again will complete the object name up to
uniqueness, etc. If you just wish to see what completions exist without
adding any extra characters, type M-?.
M-? List all possible completions of the object name at point.
ESS also provides completion over the components of named lists and environments (after ‘$’), S4 classes slots (after @), package and namespace objects (after :: and :::).
Completion is also provided over file names, which is particularly
useful when using R functions such as get() or scan()
which require fully expanded file names.
In the iESS buffer, if the cursor is not in a string and does not follow a (partial) object name, the TAB key has a third use: it expands history references. See References to historical commands.
Efficiency in completion is gained by maintaining a cache of objects currently known to R; when a new object becomes available or is deleted, only one component of the cache corresponding to the associated directory needs to be refreshed. If ESS ever becomes confused about what objects are available for completion (such as when if refuses to complete an object you know is there), the command M-x ess-resynch forces the entire cache to be refreshed, which should fix the problem.
When inside a function call (i.e. following ‘(’), TAB
completion also provides function arguments. If function is a generic,
completion will provide all the arguments of S3 methods known to
R.
In R versions newer than 2.14.1, ESS uses R’s built-in completion
engine and will honor its settings. You may read about them by issuing
?rc.options at the prompt. For example, to have the equals sign
for function arguments contain spaces on both sides (e.g.
lm(for) suggests formula = rather than
formula=), you can adjust funarg.suffix:
utils::rc.options(funarg.suffix = " = ")
A related functionality is See ElDoc, which displays function arguments in the echo area whenever the point is inside a function call.
ESS uses IDO mechanism (part of default
Emacs) for minibuffer completion if the ido package is available
and the value of ess-use-ido it t (the default). The
completion command ess-completing-read falls back on classic
completion-read interface if this feature is not available for
whatever reason.
Another popular package for completion is company, short for complete
anything. ESS provides support for company out-of-the-box. To disable
company support, set ess-use-company to nil. You can set
it to 'script-only to only activate company in R scripts.
Another option for comprehensively handling completion in Emacs is via Icicles (https://www.emacswiki.org/emacs/Icicles). It allows users to have completions shown temporarily in the standard ‘*Completions*’ buffer, and interactively select completion candidates using several methods. As of version 2013.04.04, Icicles provides support for completion in ESS. Please consult Icicles documentation, which is easily accessible via customize-group Icicles, for more details on installation and customization options.
Once installed, Icicles can be activated by evaluating):
(require 'icicles) (icy-mode 1)
Icicles can be toggled by typing M-x icy.
When Icicles is on, TAB offers completion, provided the conditions
determined by ess-first-tab-never-complete allow it. Typing
M-TAB will attempt completion regardless. Typing M-? in
ESS or iESS modes brings up the relevant completion
candidates from which to choose. Minibuffer input filters the available
candidates. Use TAB for prefix completion or S-TAB for
substring or regexp completion. Use S-SPC to match an additional
pattern (repeatable). You can cycle among the matching candidates,
choosing with RET or mouse-2.
ESS provides several tools to help you with the development of your R packages:
ESS tracebug offers visual debugging, interactive error
navigation, interactive backtrace, breakpoint manipulation, control over
R error actions, watch window and interactive flagging/unflagging of
functions for debugging.
ess-tracebug is on by default. You can toggle it on
and off with M-x ess-tracebug. To disable startup
activation of ess-tracebug set ess-use-tracebug to nil.
Tracebug functionality can be found on ess-dev-map, bound to
C-c C-t. Additionally, when subprocess is in a debugging state
ess-debug-minor-mode is active and the following additional
shortcuts are available:
* Interactive Debugging (`ess-debug-minor-mode-map'): M-C . Continue . `ess-debug-command-continue' M-C-C . Continue multi . `ess-debug-command-continue-multi' M-N . Next step . `ess-debug-command-next' M-C-N . Next step multi . `ess-debug-command-next-multi' M-U . Up frame . `ess-debug-command-up' M-Q . Quit debugging . `ess-debug-command-quit'
These are all the tracebug commands defined in ess-dev-map
(C-c C-t ? to show this table):
* Breakpoints (`ess-dev-map'): b . Set BP (repeat to cycle BP type) . `ess-bp-set' B . Set conditional BP . `ess-bp-set-conditional' k . Kill BP . `ess-bp-kil' K . Kill all BPs . `ess-bp-kill-all' o . Toggle BP state . `ess-bp-toggle-state' l . Set logger BP . `ess-bp-set-logger' n . Goto next BP . `ess-bp-next' p . Goto previous BP . `ess-bp-previous' (C- prefixed equivalents are also defined) * Debugging (`ess-dev-map'): ` . Show traceback . `ess-show-traceback' (also on C-c `) ~ . Show callstack . `ess-show-call-stack' (also on C-c ~) e . Toggle error action (repeat to cycle). `ess-debug-toggle-error-action' d . Flag for debugging . `ess-debug-flag-for-debugging' u . Unflag for debugging . `ess-debug-unflag-for-debugging' w . Watch window . `ess-watch' (C- prefixed equivalents are also defined) * Navigation to errors (general Emacs functionality): C-x `, M-g n . `next-error' M-g p . `previous-error' * Misc: ? . Show this help . `ess-tracebug-show-help'
To configure how electric watch window splits the display see
ess-watch-width-threshold and ess-watch-height-threshold
variables.
Note: Currently, ess-tracebug does not detect some of R’s debug related messages in non-English locales. To set your R messages to English add the following line to your .Rprofile init file:
Sys.setlocale("LC_MESSAGES", "C")
Consider a buffer with the following function:
test <- function(x){
mean(x),
}
Evaluating the function (e.g. C-c C-c) results in an error
about an unexpected comma. You can use next-error (bound by
default to C-x `, M-g n, and M-g M-n) to jump to
the place where the error occurred. Alternatively, use the mouse to
click on the error to jump to where it occurred.
Correct the error by deleting the comma. Now put point on mean
and set the breakpoint using C-c C-t b (ess-bp-set) and
reevaluate the function. Jump to the inferior buffer (possibly using
C-c C-z) and evaluate test(1:10). An interactive debug
process starts, stopping at the breakpoint we just specified. Here you
can debug your function (what is x at this point?). Use
M-N to continue.
Let’s replace our test function with one slightly more complicated:
test <- function(x){
x <- x + 1
y <- mean(x)
x <- ifelse(x > 5, x, x - 100)
list(x, y)
}
Try setting multiple breakpoints. You can unset a breakpoint by killing
it with C-c C-t k. You can set conditional breakpoints too. Try
setting one by placing point on the line x <- x + 1 and doing
C-c C-t B. ESS will ask for the condition. Let’s set it to
!is.numeric(x). After re-evaluating test, try calling
test(1:100) and test('foo').
You can remove all breakpoints with C-c C-t K.
You can flag a function for debugging (similar to calling
debug(test) at R’s prompt) by doing C-c C-t d. Try this
yourself by putting point over test and doing C-c C-t d.
If an error occurs, you can get the complete call stack by doing
C-c ` or C-c C-t ` (ess-show-traceback).
Tracebug also offers a watch window where you can watch values of
objects. Open it with C-c C-t w (ess-watch). Initially you
aren’t watching anything. Add something with a (e.g. a
test). The watch window displays what the object is at any given time
and automatically updates. Quit the watch window with q.
ESS provides two ways of writing documentation for R objects. Either using the standard R documentation system or using in-source documentation written as structured comment fields for use with the roxygen2 package.
R objects are documented in files written in the R documentation (“Rd”), a simple markup language closely resembling (La)TeX, which can be processed into a variety of formats, including LaTeX, HTML, and plain text. Rd format is described in section “Rd format” of the “Writing R Extensions” manual in the R distribution. ESS has several features that facilitate editing Rd files.
Visiting an Rd file as characterized by its extension Rd will activate Rd Mode, which provides several facilities for making editing R documentation files more convenient, by helping with indentation, insertions, even doing some of the typing for you (with Abbrev Mode), and by showing Rd keywords, strings, etc. in different faces (with Font Lock Mode).
In Rd mode, the following special Emacs commands can be used in addition to the standard Emacs commands.
Describe the features of Rd mode.
Reindent the current line, insert a newline and indent the new line
(reindent-then-newline-and-indent). An abbrev before point is
expanded if abbrev-mode is non-nil.
Indent current line based on its contents and on previous lines.
(indent-according-to-mode).
Insert a “skeleton” with Rd markup for at least all mandatory entries
in Rd files (Rd-mode-insert-skeleton). Note that many users
might prefer to use the R function prompt on an existing R object
to generate a non-empty Rd “shell” documenting the object (which
already has all information filled in which can be obtained from the
object).
Insert “font” specifiers for some of the Rd markup commands markup
available for emphasizing or quoting text, including markup for URLs and
email addresses (Rd-font). C-c C-f is only a prefix; see
e.g. C-c C-f TAB for the available bindings. Note that
currently, not all of the Rd text markup as described in section
“Marking text” of “Writing R Extensions” can be accessed via
C-c C-f.
Insert a suitably indented ‘\item{’ on the next line
(Rd-mode-insert-item).
Preview a plain text version (“help file”, see Reading help files) generated
from the Rd file (Rd-preview-help).
In addition, when editing Rd files one can interact with a running R process in a similar way as when editing R language files. For example, C-c C-v provides access to on-line help, and C-c C-n sends the current line to the R process for evaluation. This interaction is particularly useful when editing the examples in the Rd file. See C-h m for all available commands.
Rd mode also provides access to abbreviations for most of the Rd markup commands. Type M-x list-abbrevs with Abbrev mode turned on to list all available abbrevs. Note that all Rd abbrevs start with a grave accent.
Rd mode can be customized via the following variables.
Hook to be run when Rd mode is entered.
The indentation of Rd code with respect to containing blocks. Default is 2.
The shell command used for converting Rd source to help text. Default is ‘R CMD Rd2txt’.
To automatically turn on the abbrev and font-lock features of Rd mode, add the following lines to one of your Emacs startup files:
(add-hook 'Rd-mode-hook
(lambda ()
(abbrev-mode 1)
(font-lock-mode 1)))
The roxygen2 R package makes it possible to keep the intended contents for Rd files as structured comments in the R source files. roxygen2 can then parse R files and generate appropriate Rd files from these comments, fill the usage fields as well as updating NAMESPACE files. See the roxygen2 documentation found via https://roxygen2.r-lib.org for more information on its usage. An example of an roxygen2 entry for a simple R function can look like this:
##' Description of the function
##'
##' Further details about this function
##' @title A title
##' @param me all parameters must be listed and documented
##' @return Description of the return value
##' @author The author
myfun <- function(me)
cat("Hello", me, "\n")
The entry is immediately preceding the object to document and all lines
start with the roxygen2 prefix string, in this case ##'. ESS
provides support to edit these documentation entries by providing line
filling, navigation, template generation etc. Syntax highlighting is
provided.
roxygen2 is customized by the variables in the customization group “Ess Roxy”. Customizations include the roxygen2 prefix, use of folding to toggle visibility of roxygen2 entries and the roxygen2 template.
All ESS roxygen2 support is defined in ess-roxy.el which is loaded by default in ESS. The following special Emacs commands are provided.
C-c C-o C-o Generate a roxygen2 template or update the parameter
list in roxygen2 entry at point (or above the function at the
point). Documented parameters that are not in the function are placed
last in the list, parameters that are not documented and not in the
definition are dropped. Parameter descriptions are filled if
ess-roxy-fill-param-p is non-nil.
C-c C-o C-c Toggle the presence of the leading roxygen2 string on all lines in the marked region (between beg and end. Convenient for transferring text to roxygen2 entries and to evaluate example fields.
C-c C-o C-r Use the attached R process to parse the entry at point to obtain the Rd code. Convenient for previewing and checking syntax. When used with the prefix argument name-file, i.e. C-u C-c C-e C-r, place the content in a buffer associated with a Rd file with the same name as the documentation. Requires the roxygen2 package to be installed.
C-c C-o C-t Use the attached R process to parse the entry at to generate HTML for the entry and open it in a browser. When used with the prefix argument visit-instead-of-open, i.e. C-u C-c C-e C-t, visit the generated HTML file instead. Requires the roxygen2 and tools packages to be installed.
C-c C-o p Go to start of the roxygen2 entry above point.
C-c C-o n Go to end of the roxygen2 entry above point.
C-c C-o C-h Use the hideshow mode to fold away the visibility of all roxygen2 entries. Hide-show support must be enabled for this binding to get defined.
ESS also advises the following standard editing functions in order to make roxygen2 editing more intuitive:
ess-R-complete-object-name Complete roxygen2 tag at
point. E.g. doing TAB when the point is at the end of @par
completes to @param.
mark-paragraph If the transient mark mode is active, place mark
and point at start end end of the field at point and activate the mark.
ess-indent-command If hide-show support is enabled, fold away the
visibility of the roxygen2 entry at point.
fill-paragraph Fill the roxygen2 field at point.
ess-roxy-move-beginning-of-line Move to the point directly to the right of
the roxygen2 start string. If already there, move to the beginning of the line.
newline-and-indent Insert a new line and the roxygen2 prefix
string.
In non package files evaluation commands, See Sending code to the ESS process, send
portions of the current buffer environment (R_GlobalEnv. When
developing packages, ESS sends code to the corresponding package
namespace and (for visible objects) into package environment (visible on
search path). All objects that are assigned are displayed in the
minibuffer alongside the environment in which they are assigned.
Here is a short overview of how namespace and package environments work in R. All objects defined in a package ’foo’ are stored in an environment called ’namespace:foo’. Parent environment of ’namespace:foo’ is an environment ’imports:foo’ which contains copies of all objects from other packages which ’foo’ imports. Parent environment of ’imports:foo’ is the ’namespace:base’. Parent environment of ’namespace:base’ is .GlobalEnv. Thus functions and methods stored in ’namespace:foo’ see all the objects in .GlobalEnv unless shadowed by objects in ’imports:foo’, ’namespace:base’, or ’namespace:foo’ itself. There is another environment associated with ’foo’ - ’package:foo’. This environment stores *copies* of exported objects from ’namespace:foo’ and is placed on the search() path, i.e. if ’foo’ is loaded and if you start with .GlobalEnv and iteratively call parent.env() you will get eventually to ’package:foo’. Thus all methods and functions defined in .GlobalEnv can "see" objects in ’package:foo’ environment. See also https://cran.r-project.org/doc/manuals/R-ints.html#Namespaces.
Occasionally you want to evaluate into a package from a non-package file,
or the other way around, evaluate into GlobalEnv from inside a
package. In such cases C-c C-t C-s is your friend.
C-c C-t C-s Set or unset the current evaluation environment (a package).
ESS has a few extra features, which didn’t fit anywhere else.
In ElDoc mode, the echo area displays function’s arguments at
point. ElDoc is active by default in ess-mode and
inferior-ess-mode buffers. To activate it only in ess-mode
buffers, place the following into your init file:
(setq ess-use-eldoc 'script-only)
If ‘t’, activate eldoc in ess-mode and inferior-ess-mode buffers.
If ‘'script-only’ activate in ess-mode buffers only. Set
ess-use-eldoc to nil to stop using ElDoc
altogether.
This variable controls whether the help is shown only inside function
calls. If set to ‘t’, ElDoc shows help string whenever the
point is on a symbol, otherwise (the default), shows only when the point
is in a function call, i.e. after ‘'('’.
The variable determines how the doc string should be abbreviated to fit into minibuffer. Posible values are ‘nil’, ‘mild’, ‘normal’, ‘strong’ and ‘aggressive’. Please see the documentation of the variable for more details. The default filter is ‘normal’.
Ess-eldoc also honors the value of
eldoc-echo-area-use-multiline-p, which if set to ‘nil’, will
cause the truncation of doc string indifferent of the value of
ess-eldoc-abbreviation-style. This way you can combine different
filter levels with the truncation.
The minor mode flymake-mode provides on-the-fly syntax
checking. ESS provides support for flymake in R-mode in Emacs versions
26 and newer. It is enabled by default, to disable it you may set
ess-use-flymake to nil. In order to use it, you may need
to install the lintr R package, available from CRAN.
When non-nil, use flymake. If 'process, only use flymake when the
buffer has an inferior process running.
This variable describes the linters to use. It can either be a string
with an R expression to be used as-is or a list of strings where each
element is passed to lintr::with_defaults. See the help page
for lintr::default_linters for information on available linters
and their defaults.
When t (the default), lintr uses a cache. Change to nil
to disable lintr’s caching mechanism.
Request and execute a command from ess-handy-commands list.
An alist of custom ESS commands available for call by
ess-handy-commands and ess-smart-comma function.
Currently contains:
ess-change-directory
ess-display-index
ess-display-help-on-object
ess-display-vignettes
ess-execute-objects
ess-execute-search
ess-execute-screen-options
ess-install.packages
ess-library
ess-setRepositories
ess-sos
Handy commands: ess-library, ess-install.packages, etc -
ask for item with completion and execute the correspond
command. ess-sos is a interface to findFn function in
package sos. If package sos is not found, ask user for
interactive install.
ESS provides Font-Lock (see Using Multiple Typefaces in The Gnu Emacs Reference Manual) patterns for inferior R Mode, S Mode, and R Transcript Mode buffers.
The font-lock patterns are defined by the following variables, which you may modify if desired:
Font-lock patterns for inferior *R* processes. (There is a
corresponding inferior-S-font-lock-keywords for *S* processes.)
The default value highlights prompts, inputs, assignments, output
messages, vector and matrix labels, and literals such as ‘NA’ and
TRUE.
Font-lock patterns for ESS R programming mode. (There is a
corresponding ess-S-font-lock-keywords for S buffers.) The
default value highlights function names, literals, assignments, source
functions and reserved words.
Emacs has facilities for highlighting the parenthesis matching the parenthesis at point. This feature is very useful when trying to examine which parentheses match each other. This highlighting also indicates when parentheses are not matching. You may activate it putting this in your Emacs configuration file:
(show-paren-mode)
One of the main features of R is its ability to generate high-resolution graphics plots. When using R in a windowing environment, no additional configuration is necessary; plots will be shown in a new (non-Emacs) window.
If not using a windowing environment or S, ESS provides a number of features for dealing with such plots.
printer() driverThis is the simplest (and least desirable) method of using graphics
within ESS. S’s printer() device driver produces crude character
based plots which can be contained within the ESS process buffer
itself. To start using character graphics, issue the S command
printer(width=79)
(the width=79 argument prevents Emacs line-wrapping at column
80 on an 80-column terminal. Use a different value for a terminal with
a different number of columns.) Plotting commands do not generate
graphics immediately, but are stored until the show() command
is issued, which displays the current figure.
Of course, the ideal way to use graphics with ESS is to use a windowing system. Under X Windows, or X11, this requires that the DISPLAY environment variable be appropriately set.
S+6.1 and newer on Windows contains a java library that supports graphics. Send the commands:
library(winjava) java.graph()
to start the graphics driver. This allows you to use ESS for both interaction and graphics within S-PLUS. (Thanks to Tim Hesterberg for this information.)
The R and S editing modes have support for a toolbar. This toolbar provides icons to act as shortcuts for starting new S/R processes, or for evaluating regions of your source buffers. The toolbar should be present if your Emacs can display images. See Customizing ESS, for ways to change the toolbar.
Xref is an Emacs interface that supports finding “identifiers,” usually function definitions in ESS’s view. ESS ships with support for Xref in Emacs versions 25.1 and higher. For how to use this feature, see See Xref in The Gnu Emacs Reference Manual.
Ess-rdired provides a dired-like buffer for viewing, editing and plotting objects in your current R session. If you are used to using the dired (directory editor) facility in Emacs, this mode gives you similar functionality for R objects.
Start an R session with M-x R and then store a few variables, such as:
s <- sin(seq(from=0, to=8*pi, length=100)) x <- c(1, 4, 9) y <- rnorm(20) z <- TRUE
Then use M-x ess-rdired to create a buffer listing the objects in your current environment and display it in a new window:
mode length s numeric 100 x numeric 3 y numeric 20 z logical 1
Type C-h m or ? to get a list of the keybindings for this mode. For example, with your point on the line of a variable, ‘p’ will plot the object, ‘v’ will view it, and ‘d’ will mark the object for deletion (‘x’ will actually perform the deletion).
ESS[R] provides several commands to list and manage packages and
objects under the C-c C-. keymap:
C-c C-. l List all packages in all available libraries.
C-c C-. r List available packages from repositories listed by
getOptions(``repos'') in the current R session.
C-c C-. u Update packages in a particular library lib and repository repos.
C-c C-. m Remove all R objects.
C-c C-. w Load a workspace file into R.
C-c C-. s Save a workspace file.
C-c C-. d Change the working directory for the current R session.
C-c C-. H Use browse-url to navigate R html
documentation.
Functions for listing objects and packages
(ess-rutils-local-pkgs, ess-rutils-repos-pkgs, and
ess-rutils-objs) show results in a separate buffer and window,
in ess-rutils-mode, providing useful key bindings in this mode
(type ? in this buffer for a description).
Org-mode (https://orgmode.org) now supports reproducible research and literate programming in many languages (including R) – see chapter 15 of the Org manual (https://orgmode.org/org.html#Working-with-Source-Code). For ESS users, this offers a document-based work environment within which to embed ESS usage. R code lives in code blocks of an Org document, from which it can be edited in ess-mode, evaluated, extracted ("tangled") to pure code files. The code can also be exported ("woven") with the surrounding text to several formats including HTML and LaTeX. Results of evaluation including figures can be captured in the Org document, and data can be passed from the Org document (e.g. from a table) to the ESS R process. (This section contributed by Dan Davison and Eric Schulte.)
Libraries for literate data analysis are obsolete and not loaded by
default. This includes ess-noweb, ess-swv, and related
functionality like Rnw-mode. Users are encouraged to switch to
one of several other packages that deal with these modes. For example,
polymode https://github.com/polymode/poly-R/,
https://polymode.github.io/, or markdown-mode with edit-indirect
https://jblevins.org/projects/markdown-mode.
ESS provides support for writing and processing Sweave (https://leisch.userweb.mwn.de/Sweave/), building up on Emacs’ ess-noweb-mode for literate programming. When working on an Sweave document, the following key bindings are available:
M-n s Run Sweave on the current .Rnw file. If choose is non-‘nil’, offer a menu of available weavers.
M-n l Run LaTeX after Sweave’ing.
M-n p Generate and display a postscript file after LaTeX’ing.
M-n P Generate and display a PDF file after LaTeX’ing. Optional
argument pdflatex-cmd is the command to use, which by default, is
the command used to generate the PDF file is the first element of
ess-swv-pdflatex-commands.
Commands used by ess-swv-PDF to run a version of pdflatex; the
first entry is the default command.
Sweave’ing with ess-swv-weave starts an inferior-ESS process, if
one is not available. Other commands are available from the
‘Sweaving, Tangling, ...’ submenu of the Noweb menu.
AUCTeX (https://www.gnu.org/software/auctex) users may prefer to
set the variable ess-swv-plug-into-AUCTeX-p (available from the
“ESS Sweave” customization group) to t. Alternatively, the same can
be achieved by activating the entry “AUCTeX Interface” from the
‘Sweaving, Tangling, ...’ submenu, which toggles this variable on
or off. When the interface is activated, new entries for Sweave’ing and
LaTeX’ing thereafter are available from AUCTeX’s “Command” menu.
Sweave’ing can, thus, be done by C-c C-c Sweave RET without an
inferior-ESS process. Similarly, LaTeX’ing can be done by C-c C-c
LaTeXSweave RET. In both cases, the process can be monitored with
C-c C-l (TeX-recenter-output-buffer). Open the viewer with
C-c C-v (TeX-view), as usual in AUCTeX.
ESS[R] mode should be automatically turned on when visiting a file ending with an R or S suffix (*.R, *.S, *.s, etc), which enables the features discussed previously. Alternatively, type M-x R-mode to put the current buffer into R mode. However, one will have to start up an inferior process to take advantage of the interactive features.
To start up iESS mode for R, use:
M-x R M-x S+3 M-x S4 M-x S+5 M-x S+6
(for R, S-PLUS 3.x, S4, S+5, S+6 or S+7, respectively. This assumes that you have access to each).
In the case that you wish to pass command line arguments to the starting
R process, call it with the universal prefix. To set command line
arguments that apply to all future iESS sessions, set the variable
inferior-R-args.
Note that R has some extremely useful command line arguments. For
example, --vanilla will ensure R starts up without loading in any init
files.
If you have other versions of R or S available on the system, ESS
is also able to start those versions. How this exactly works depend on
which OS you are using (details below). The
general principle, regardless of OS, is that ESS searches the paths
listed in the variable exec-path for R binaries. If ESS cannot
find your R binaries, on Unix you can change the Unix environment
variable PATH, as this variable is used to set exec-path.
R on GNU/Linux systems and other Unix-like systems (macOS): If you have
"R-3.6.3" on your exec-path, it can be started using M-x
R-3.6.3. By default, ESS will find versions of R beginning “R-1”,
“R-2”, …, “R-7”, “R-devel”, or “R-patched”. If your versions of R
are called other names, consider renaming them with a symbolic link or
change the variable ess-r-runners-prefixes. To see which functions have
been created for starting different versions of R, type M-x R- and
then hit [Tab]. These other versions of R can also be started from the
"ESS->Start Process->Other" menu.
R on Windows systems: If you have "rw1081" on your exec-path, it
can be started using M-x rw1081. By default, ESS will find
versions of R located in directories parallel to the version of R in
your PATH. If your versions of R are called other names, you
will need to change the variable ess-rterm-versions. To see
which functions have been created for starting different versions of R,
type M-x rw and then hit [Tab]. These other versions of
R can also be started from the "ESS->Start Process->Other" menu.
Once ESS has found these extra versions of R, it will then create a new
function, called M-x run-ess-r-newest, which will call the newest version
of R that it found. (ESS examines the date in the first line of
information from R --version to determine which is newest.)
There are two philosophies for using ESS. Most modern best practices prefer the first. ESS is configured for the first, and this is what the manual focuses on.
1: The source code is real. Objects are realizations of the source code. Source for EVERY user modified object is placed in a particular directory or directories, for later editing and retrieval.
2: R objects are real. Source code is a temporary realization of the objects. Dumped buffers should not be saved. _We_strongly_discourage_this_approach_. However, if you insist, add the following lines to your Emacs configuration file:
(setq ess-keep-dump-files nil) (setq ess-delete-dump-files t) (setq ess-mode-silently-save nil)
Comments as to what should be happening are prefixed by "##".
### Data Analysis Example
## Open the file you want with
C-x C-f myfile.R
## Edit as appropriate, then start and switch to an *R*
process buffer with
C-c C-z
## A new buffer *R* will appear. R will have been started
## in this buffer. The buffer is in iESS [R] mode.
## Return to the script
C-z (or C-c C-z if you have pressed other keys in between)
## Send current region (if active), or function, or paragraph at
point with
C-M-x
## Send current region or function or paragraph and step with
C-c C-c
## Send current line with
C-RET
## or load the entire buffer with
C-c C-l
## Save the revised myfile.R with
C-x C-s.
## Save the entire *R* interaction buffer with
C-x C-s.
## Search for documentation with
C-c C-d
## Investigate available ESS commands with
C-c C-h
## You will be prompted for a file name. The recommended name is
## myfile.Rout. With this suffix, the file will be opened in
## ess-transcript.
ESS[SAS] was designed for use with SAS. It is descended from emacs
macros developed by John Sall for editing SAS programs and SAS-mode by
Tom Cook. Those editing features and new advanced features are part of
ESS[SAS]. The user interface of ESS[SAS] has similarities with ESS[S]
and the SAS Display Manager.
ESS[SAS] was designed to aid the user in writing and maintaining SAS programs, such as foo.sas. Both interactive and batch submission of SAS programs is supported.
ESS[SAS] was written with two primary goals.
ESS[SAS] is the mode for editing SAS language files. This mode handles:
ESS[SAS] is automatically turned on when editing a file with a .sas
suffix (or other extension, if specified via auto-mode-alist). The function
keys can be enabled to use the same function keys that
the SAS Display Manager does. The interactive capabilities of ESS require you
to start an inferior SAS process with M-x SAS
(See iESS[SAS]–Interactive SAS processes.)
At this writing, the indenting and syntax highlighting are generally
correct. Known issues: for multiple line * or %* comments, only the
first line is highlighted; for .log files, only the first line of a
NOTE:, WARNING: or ERROR: message is highlighted; unmatched
single/double quotes in CARDS data lines are NOT ignored; in an
iterative DO statement, TO and BY are not highlighted.
Two options. The TAB key is bound by default to
sas-indent-line. This function is used to syntactically indent
SAS code so PROC and RUN are in the left margin,
other statements are indented sas-indent-width spaces from the
margin, continuation lines are indented sas-indent-width spaces
in from the beginning column of that statement. This is the type of
functionality that emacs provides in most programming language modes.
This functionality is activated by placing the following line in your
initialization file prior to a require/load:
(setq ess-sas-edit-keys-toggle nil)
ESS provides an alternate behavior for TAB that makes it
behave as it does in SAS Display Manager, i.e. move the cursor to
the next stop. The alternate behavior also provides a "TAB"
backwards, C-TAB, that moves the cursor to the stop to
the left and deletes any characters between them. This functionality
is obtained by placing the following line in your initialization file
prior to a require/load:
(setq ess-sas-edit-keys-toggle t)
Under the alternate behavior, TAB is bound to M-x tab-to-tab-stop
and the stops are defined by ess-sas-tab-stop-list.
Submission of a SAS batch job is dependent on your environment.
ess-sas-submit-method is determined by your operating system and
your shell. It defaults to 'sh unless you are running Windows or
Mac Classic. Under Windows, it will default to 'sh if you are using
a UNIX-imitating shell; otherwise 'ms-dos for an MS-DOS
shell. On macOS, it will default to 'sh, but under Mac Classic,
it defaults to 'apple-script. You will also set this to 'sh
if the SAS batch job needs to run on a remote machine rather than your
local machine. This works transparently if you are editing the remote file
via ange-ftp/EFS or tramp. Note that
ess-sas-shell-buffer-remote-init is a Local Variable that defaults
to "ssh" which will be used to open the buffer on the remote host
and it is assumed that no password is necessary, i.e. you are using
ssh-agent/ssh-add or the equivalent (see the discussion about
Local Variables below if you need to change the default).
However, if you are editing the file locally and transferring it back and
forth with Kermit, you need some additional steps. First, start Kermit
locally before remotely logging in. Open a local copy of the file with the
ess-kermit-prefix character prepended (the default is "#").
Execute the command M-x ess-kermit-get which automatically brings
the contents of the remote file into your local copy. If you transfer
files with Kermit manually in a *shell* buffer, then note that the
Kermit escape sequence is C-q C-\ c rather than C-\ c which
it would be in an ordinary terminal application, i.e. not in an emacs
buffer. Lastly, note that the remote Kermit command is specified by
ess-kermit-command.
The command used by the SUBMIT function key (F3 or F8)
to submit a batch SAS job, whether local or remote, is
ess-sas-submit-command which defaults to sas-program.
sas-program is "invoke SAS using program file" for Mac
Classic and "sas" otherwise. However, you may have to alter
ess-sas-submit-command for a particular program, so it is defined as
buffer-local. Conveniently, it can be set at the end of the program:
endsas; Local variables: ess-sas-submit-command: "sas8" End:
The command line is also made of ess-sas-submit-pre-command,
ess-sas-submit-post-command and ess-sas-submit-command-options
(the last of which is also buffer-local).
Here are some examples for your ~/.emacs or ~/.emacs.d/init.el
file (you may also use M-x customize-variable):
;'sh default (setq ess-sas-submit-pre-command "nohup") ;'sh default (setq ess-sas-submit-post-command "-rsasuser &") ;'sh example (setq-default ess-sas-submit-command "/usr/local/sas/sas") ;'ms-dos default (setq ess-sas-submit-pre-command "start") ;'ms-dos default (setq ess-sas-submit-post-command "-rsasuser -icon") ;Windows example (setq-default ess-sas-submit-command "c:/progra~1/sas/sas.exe") ;Windows example (setq-default ess-sas-submit-command "c:\\progra~1\\sas\\sas.exe")
There is a built-in delay before a batch SAS job is submitted when using a UNIX-imitating shell under Windows. This is necessary in many cases since the shell might not be ready to receive a command. This delay is currently set high enough so as not to be a problem. But, there may be cases when it needs to be set higher, or could be set much lower to speed things up. You can over-ride the default in your ~/.emacs or ~/.emacs.d/init.el file by:
(setq ess-sleep-for 0.2)
For example, (setq ess-sas-global-unix-keys t) keys shown,
(setq ess-sas-global-pc-keys t) in parentheses;
ESS[SAS] function keys are presented in the next section.
Open the file you want to work with C-x C-f foo.sas.
foo.sas will be in ESS[SAS] mode. Edit as appropriate, then save and
submit the batch SAS job.
F3 (F8)
The job runs in the *shell* buffer while you continue to edit
foo.sas. If ess-sas-submit-method is 'sh, then the
message buffer will display the shell notification when the
job is complete. The 'sh setting also allows you to
terminate the SAS batch job before it is finished.
F8 (F3)
Terminating a SAS batch in the *shell* buffer.
kill PID
You may want to visit the .log (whether the job is still running or it is finished) and check for error messages. The .log will be refreshed and you will be placed in it’s buffer. You will be taken to the first error message, if any.
F5 (F6)
Goto the next error message, if any.
F5 (F6)
Now, ‘refresh’ the .lst and go to it’s buffer.
F6 (F7)
If you wish to make changes, go to the .sas file with.
F4 (F5)
Make your editing changes and submit again.
F3 (F8)
The setup of function keys for SAS batch processing is unavoidably complex, but the usage of function keys is simple. There are five distinct options:
Option 1 (default). Function keys in ESS[SAS] are not bound to Elisp commands. This is in accordance with the GNU Elisp Coding Standards (GECS) which do not allow function keys to be bound so that they are available to the user.
Options 2-5. Since GECS does not allow function keys to be bound by modes, these keys are often unused. So, ESS[SAS] provides users with the option of binding Elisp commands to these keys. Users who are familiar with SAS will, most likely, want to duplicate the function key capabilities of the SAS Display Manager. There are four options (noted in parentheses).
auto-mode-alist.
The distinction between local and global is subtle. If you want the ESS[SAS] definitions to work when you are in the *shell* buffer or when editing files other than the file extensions that ESS[SAS] recognizes, you will most likely want to use the global definitions. If you want your function keys to understand SAS batch commands when you are editing SAS files, and to behave normally when editing other files, then you will choose the local definitions. The option can be chosen by the person installing ESS for a site or by an individual.
require/load. ESS[SAS] function keys are available
in ESS[SAS] if you choose either 2 or 3 and in all modes if you
choose 4 or 5:
;;2; (setq ess-sas-local-unix-keys t) ;;3; (setq ess-sas-local-pc-keys t) ;;4; (setq ess-sas-global-unix-keys t) ;;5; (setq ess-sas-global-pc-keys t)
The names -unix- and -pc- have nothing to do with the operating system
that you are running. Rather, they mimic the definitions that the SAS
Display Manager uses by default on those platforms.
(load "ess-site") ;; local-unix-keys (ess-sas-global-pc-keys)
Finally, we get to what the function keys actually do. You may recognize some of the nicknames as SAS Display Manager commands (they are in all capitals).
| UNIX | PC | Nickname |
| F2 | F2 | ‘refresh’ |
| revert the current buffer with the file of the same name if the file is newer than the buffer | ||
| F3 | F8 | SUBMIT |
| save the current .sas file (which is either the .sas file in the current buffer or the .sas file associated with the .lst or .log file in the current buffer) and submit the file as a batch SAS job | ||
| F4 | F5 | PROGRAM |
| switch buffer to .sas file | ||
| F5 | F6 | LOG |
| switch buffer to .log file, ‘refresh’ and goto next error message, if any | ||
| F6 | F7 | OUTPUT |
| switch buffer to .lst file and ‘refresh’ | ||
| F7 | F4 | ‘filetype-1’ |
| switch buffer to ‘filetype-1’ (defaults to .txt) file and ‘refresh’ | ||
| F8 | F3 | ‘shell’ |
| switch buffer to *shell* | ||
| F9 | F9 | VIEWTABLE |
open an interactive PROC FSEDIT session on the SAS
dataset near point | ||
| F10 | F10 | toggle-log |
| toggle ESS[SAS] for .log files; useful for certain debugging situations | ||
| F11 | F11 | ‘filetype-2’ |
| switch buffer to ‘filetype-2’ (defaults to .dat) file and ‘refresh’ | ||
| F12 | F12 | viewgraph |
open a GSASFILE near point for viewing either in emacs
or with an external viewer | ||
| C-F1 | C-F1 | rtf-portrait |
| create an MS RTF portrait file from the current buffer with a file extension of .rtf | ||
| C-F2 | C-F2 | rtf-landscape |
| create an MS RTF landscape file from the current buffer with a file extension of .rtf | ||
| C-F3 | C-F8 | submit-region |
| write region to ess-temp.sas and submit | ||
| C-F5 | C-F6 | append-to-log |
| append ess-temp.log to the current .log file | ||
| C-F6 | C-F7 | append-to-output |
| append ess-temp.lst to the current .lst file | ||
| C-F9 | C-F9 | INSIGHT |
open an interactive PROC INSIGHT session on the SAS
dataset near point | ||
| C-F10 | C-F10 | toggle-listing |
| toggle ESS[SAS] for .lst files; useful for toggling read-only |
SUBMIT, PROGRAM, LOG and OUTPUT need no further
explanation since they mimic the SAS Display Manager commands and
related function key definitions. However, six other keys have been
provided for convenience and are described below.
‘shell’ switches you to the *shell* buffer where you can
interact with your operating system. This is especially helpful if you
would like to kill a SAS batch job. You can specify a different buffer
name to associate with a SAS batch job (besides *shell*) with the
buffer-local variable ess-sas-shell-buffer. This allows you to have
multiple buffers running SAS batch jobs on multiple local/remote
computers that may rely on different methods specified by the buffer-local
variable ess-sas-submit-method.
F2 performs the ‘refresh’ operation on the current buffer.
‘refresh’ compares the buffer’s last modified date/time with the
file’s last modified date/time and replaces the buffer with the file if the
file is newer. This is the same operation that is automatically performed
when LOG, OUTPUT, ‘filetype-1’ or F11 are pressed.
‘filetype-1’ switches you to a file with the same file name as your .sas file, but with a different extension (.txt by default) and performs ‘refresh’. You can over-ride the default extension; for example in your ~/.emacs or ~/.emacs.d/init.el file:
(setq ess-sas-suffix-1 "csv") ; for example
F9 will prompt you for the name of a permanent SAS dataset near
point to be opened for viewing by PROC FSEDIT. You can control the
SAS batch command-line with ess-sas-data-view-submit-options.
For controlling the SAS batch commands, you have the global variables
ess-sas-data-view-libname and
ess-sas-data-view-fsview-command as well as the buffer-local
variable ess-sas-data-view-fsview-statement. If you have your
SAS LIBNAME defined in ~/autoexec.sas, then the defaults
for these variables should be sufficient.
Similarly, C-F9 will prompt you for the name of a permanent
SAS dataset near point to be opened for viewing by PROC INSIGHT.
You can control the SAS batch command-line with
ess-sas-data-view-submit-options. For controlling the SAS batch
commands, you have the global variables ess-sas-data-view-libname
and ess-sas-data-view-insight-command as well as the buffer-local
variable ess-sas-data-view-insight-statement.
F10 toggles ESS[SAS] mode for .log files which is off by default
(technically, it is SAS-log-mode, but it looks the same). The syntax
highlighting can be helpful in certain debugging situations, but large
.log files may take a long time to highlight.
F11 is the same as ‘filetype-1’ except it is .dat by default.
F12 will prompt you for the name of a GSASFILE near the point in .log to
be opened for viewing either with emacs or with an external viewer.
Depending on your version of emacs and the operating system you are
using, emacs may support .gif and .jpg files internally. You may need
to change the following variables for your own situation.
ess-sas-graph-view-suffix-regexp is a regular expression of supported
file types defined via file name extensions.
ess-sas-graph-view-viewer-default is the default external viewer for
your platform. ess-sas-graph-view-viewer-alist is an alist of
exceptions to the default; i.e. file types and their associated
viewers which will be used rather than the default viewer.
(setq ess-sas-graph-view-suffix-regexp (concat "[.]\\([eE]?[pP][sS]\\|"
"[pP][dD][fF]\\|[gG][iI][fF]\\|[jJ][pP][eE]?[gG]\\|"
"[tT][iI][fF][fF]?\\)")) ;; default
(setq ess-sas-graph-view-viewer-default "kodakimg") ;; Windows default
(setq ess-sas-graph-view-viewer-default "sdtimage") ;; Solaris default
(setq ess-sas-graph-view-viewer-alist
'(("[eE]?[pP][sS]" . "gv") ("[pP][dD][fF]" . "gv")) ;; default w/ gv
C-F2 produces US landscape by default, however, it can produce A4 landscape (first line for "global" key mapping, second for "local"):
(global-set-key [(control f2)] 'ess-sas-rtf-a4-landscape) (define-key sas-mode-local-map [(control f2)] 'ess-sas-rtf-a4-landscape)
Inferior ESS (iESS) is the method for interfacing with interactive
statistical processes (programs). iESS[SAS] is what is needed for
interactive SAS programming. iESS[SAS] works best
with the SAS command-line option settings
"-stdio -linesize 80 -noovp -nosyntaxcheck"
(the default of inferior-SAS-args).
-stdiorequired to make the redirection of stdio work-linesize 80keeps output lines from folding on standard terminals-noovpprevents error messages from printing 3 times-nosyntaxcheckpermits recovery after syntax errors
To start up iESS[SAS] mode, use:
M-x SAS
The *SAS:1.log* buffer in ESStr mode corresponds to the file
foo.log in SAS batch usage and to the ‘SAS: LOG’ window in the
SAS Display Manager. All commands submitted to SAS, informative
messages, warnings, and errors appear here.
The *SAS:1.lst* buffer in ESSlst mode corresponds to the file
foo.lst in SAS batch usage and to the ‘SAS: OUTPUT’ window in the
SAS Display Manager. All printed output appears in this window.
The *SAS:1* buffer exists solely as a communications buffer.
The user should never use this buffer directly.
Files are edited in the foo.sas buffer. The C-c C-r key in
ESS[SAS] is the functional equivalent of bringing a file into the
‘SAS: PROGRAM EDITOR’ window followed by SUBMIT.
For example, open the file you want to work with.
C-x C-f foo.sas
foo.sas will be in ESS[SAS] mode. Edit as appropriate, and then start up SAS with the cursor in the foo.sas buffer.
M-x SAS
Four buffers will appear on screen:
| Buffer | Mode | Description |
| foo.sas | ESS[SAS] | your source file |
| *SAS:1* | iESS[SAS:1] | iESS communication buffer |
| *SAS:1.log* | Shell ESStr [] | SAS log information |
| *SAS:1.lst* | Shell ESSlst [] | SAS listing information |
If you would prefer each of the four buffers to appear in its own individual frame, you can arrange for that. Place the cursor in the buffer displaying foo.sas. Enter the sequence C-c C-w. The cursor will normally be in buffer foo.sas. If not, put it there and C-x b foo.sas.
Send regions, lines, or the entire file contents to SAS
(regions are most useful: a highlighted region will normally
begin with the keywords DATA or PROC and end with
RUN;), C-c C-r.
Information appears in the log buffer, analysis results in the listing buffer. In case of errors, make the corrections in the foo.sas buffer and resubmit with another C-c C-r.
At the end of the session you may save the log and listing buffers with the usual C-x C-s commands. You will be prompted for a file name. Typically, the names foo.log and foo.lst will be used. You will almost certainly want to edit the saved files before including them in a report. The files are read-only by default. You can make them writable by the emacs command C-x C-q.
At the end of the session, the input file foo.sas will typically have been revised. You can save it. It can be used later as the beginning of another iESS[SAS] session. It can also be used as a batch input file to SAS.
The *SAS:1* buffer is strictly for ESS use. The user should never need to read it or write to it. Refer to the .lst and .log buffers for monitoring output!
Troubleshooting: See iESS[SAS]–Common problems.
-stdio option.
SAS does not support the -stdio option on Windows.
(set-file-modes "ess-sas-sh-command" 493).
sas isn’t in your executable path; to verify, type
M-: and at the minibuffer prompt ‘Eval:’, type
(executable-find "sas")
sas
on your system calls a shell script. In that case you will need to locate
the real sas executable and link to it.
You can execute the UNIX command:
find / -name sas -print
Now place a soft link to the real sas executable in your ~/bin
directory, with for example
cd ~/bin ln -s /usr/local/sas9/sas sas
Check your PATH environment variable to confirm that
~/bin appears before the directory in which the sas
shell script appears.
Output from a SAS/GRAPH PROC can be displayed in a SAS/GRAPH
window for SAS batch on Windows or for both SAS batch and interactive
with XWindows on UNIX. If you need to create graphics files and view
them with F12, then include the following (either in
foo.sas or in ~/autoexec.sas):
filename gsasfile 'graphics.ps'; goptions device=ps gsfname=gsasfile gsfmode=append;
PROC PLOT graphs can be viewed in the listing buffer. You may
wish to control the vertical spacing to allow the entire plot
to be visible on screen, for example:
proc plot;
plot a*b / vpos=25;
run;
Or you can mark regions with the mouse and submit the code with ‘submit-region’ or paste them into SAS Display Manager.
ESS[BUGS] provides 5 features. First, BUGS syntax is described to allow for proper fontification of statements, distributions, functions, commands and comments in BUGS model files, command files and log files. Second, ESS creates templates for the command file from the model file so that a BUGS batch process can be defined by a single file. Third, ESS provides a BUGS batch script that allows ESS to set BUGS batch parameters. Fourth, key sequences are defined to create a command file and submit a BUGS batch process. Lastly, interactive submission of BUGS commands is also supported.
Model files with the .bug extension are edited in ESS[BUGS]
mode. Three keys are bound for your use in ESS[BUGS], F2,
C-c C-c and =. F2 performs the same action as it does
in ESS[SAS], See ESS[SAS]–Function keys for batch processing.
C-c C-c performs the function ess-bugs-next-action which
you will use a lot. Pressing it in an empty buffer for a model file
will produce a template for you. = inserts the set operator,
<-.
Files ending in .bmd are used for BUGS command files. When you have finished editing your model file and press C-c C-c, a command file is created if one does not already exist. When you are finished editing your command file, pressing C-c C-c again will submit your command file as a batch job.
The .bog extension is used for BUGS log files. The command line generated by ESS creates the .bog transcript file.
ESS[JAGS] provides 4 features. First, JAGS syntax is described to allow for proper fontification of statements, distributions, functions, commands and comments in JAGS model files, command files and log files. Second, ESS creates templates for the command file from the model file so that a JAGS batch process can be defined by a single file. Third, ESS provides a JAGS batch script that allows ESS to set JAGS batch parameters. Fourth, key sequences are defined to create a command file and submit a JAGS batch process.
Files with the .jag extension are edited in ESS[JAGS] mode.
Three keys are bound for your use in ESS[JAGS], F2, C-c C-c and =.
F2 performs the same action as it does in ESS[SAS],
See ESS[SAS]–Function keys for batch processing. C-c C-c performs the
function ess-bugs-next-action which you will use a lot. Pressing it
in an empty buffer for a model file will produce a template for you. =
inserts the set operator, <-.
The first press of C-c C-c will set up a template, including some
necessary file-local variables in an empty buffer. These variables are
ess-jags-chains, ess-jags-monitor, ess-jags-thin,
ess-jags-burnin and ess-jags-update; they appear in the
Local Variables section. When you are finished editing your
model file, pressing C-c C-c will perform the necessary actions to
build your command file for you.
The ess-jags-chains variable is the number of chains that
you want to initialize and sample from; defaults to 1.
The ess-jags-monitor variable is a list of variables
that you want monitored: encase each variable in double quotes.
When you press C-c C-c, the
appropriate statements are created in the command file to monitor the
list of variables. By default, no variables are explicitly monitored
which means JAGS will implicitly monitor all “default” variables.
The ess-jags-thin variable is the thinning parameter.
By default, the thinning parameter is set to 1, i.e. no thinning.
The ess-jags-burnin variable is the number of initial samples
to discard. By default, the burnin parameter is set to 10000.
The ess-jags-update variable is the number of post-burnin samples
to keep. By default, the update parameter is set to 10000. Both
ess-jags-burnin and ess-jags-update are multiplied by
ess-jags-thin since JAGS does not do it automatically.
Files ending in .jmd are for JAGS command files. For your
.jmd file, there is only one variable, ess-jags-command,
in the Local Variables section. When you have
finished editing your model file and press C-c C-c, a command
file is created if one does not already exist.
When you are finished editing your command file, pressing C-c C-c
again will submit your command file as a batch job.
The ess-jags-command variable allows you to specify a different
JAGS program to use to run your model; defaults to “jags”.
The .jog extension is used for JAGS log files. You may find F2 useful to refresh the .jog if the batch process over-writes or appends it.
ess-display-help-on-object and list completion
cannot be used while the user is entering a multi-line command. The
only real fix in this situation is to use another ESS process.
ess-eval- commands can leave point in the ESS process
buffer in the wrong place when point is at the same position as the last
process output. This proves difficult to fix, in general, as we need to
consider all windows with window-point at the right place.
ess-keep-dump-files is nil.
ess-execute causes ESS
to hang.
options(continue = " ")) are not automatically detected by ESS.
Hence, for now, the best thing is not to change the continuation prompt.
If you do change the continuation prompt, you will need to change
accordingly the value of inferior-ess-secondary-prompt in
R-customize-alist.
Please send bug reports, suggestions etc. to ESS-bugs@r-project.org, or post them on our github issue tracker
The easiest way to do this is within Emacs by typing
M-x ess-submit-bug-report
This also gives the maintainers valuable information about your installation which may help us to identify or even fix the bug.
If Emacs reports an error, backtraces can help us debug the problem. Type "M-x set-variable RET debug-on-error RET t RET". Then run the command that causes the error and you should see a *Backtrace* buffer containing debug information; send us that buffer.
Note that comments, suggestions, words of praise and large cash donations are also more than welcome.
There is a mailing list for discussions and announcements relating to ESS. Join the list by sending an e-mail with "subscribe ess-help" (or "help") in the body to ess-help-request@r-project.org; contributions to the list may be mailed to ess-help@r-project.org. Rest assured, this is a fairly low-volume mailing list.
The purposes of the mailing list include
Emacs is a complex editor with many abilities that we do not have space to describe here. If you have problems with other features of Emacs (e.g. printing), there are several sources to consult, including the Emacs FAQs (try M-x view-emacs-faq) and EmacsWiki (https://www.emacswiki.org). Please consult them before asking on the mailing list about issues that are not specific to ESS.
ESS can be easily customized to your taste simply by including the appropriate lines in your Emacs configuration file. There are numerous variables which affect the behavior of ESS in certain situations which can be modified to your liking. Keybindings may be set or changed to your preferences, and for per-buffer customizations hooks are also available.
Most of these variables can be viewed and set using the Custom facility within Emacs. Type M-x customize-group RET ess RET to see all the ESS variables that can be customized. Variables are grouped by subject to make it easy to find related variables.
| Jump to: | ,
{
}
C E M R T |
|---|
| Jump to: | ,
{
}
C E M R T |
|---|
| Jump to: | B C D E F G I M N O P Q R S |
|---|
| Jump to: | B C D E F G I M N O P Q R S |
|---|
| Jump to: | C E I R |
|---|
| Jump to: | C E I R |
|---|
| Jump to: | A B C D E F G H I K L M N O P Q R S T U W X |
|---|
| Jump to: | A B C D E F G H I K L M N O P Q R S T U W X |
|---|