Title: | R Commander |
---|---|
Description: | A platform-independent basic-statistics GUI (graphical user interface) for R, based on the tcltk package. |
Authors: | John Fox [aut, cre], Milan Bouchet-Valat [aut], Manuel Munoz Marquez [aut], Liviu Andronic [ctb], Michael Ash [ctb], Theophilius Boye [ctb], Stefano Calza [ctb], Andy Chang [ctb], Vilmantas Gegzna [ctb], Philippe Grosjean [ctb], Richard Heiberger [ctb], Yoshinobu Kanda [ctb], Kosar Karimi Pour [ctb], G. Jay Kerns [ctb], Renaud Lancelot [ctb], Matthieu Lesnoff [ctb], Uwe Ligges [ctb], Samir Messad [ctb], Martin Maechler [ctb], Robert Muenchen [ctb], Duncan Murdoch [ctb], Erich Neuwirth [ctb], Dan Putler [ctb], Brian Ripley [ctb], Miroslav Ristic [ctb], Peter Wolf [ctb], Kevin Wright [ctb] |
Maintainer: | John Fox <[email protected]> |
License: | GPL (>= 2) |
Version: | 2.9-5 |
Built: | 2024-12-24 06:35:58 UTC |
Source: | CRAN |
A platform-independent basic-statistics GUI (graphical user interface) for R, based on the tcltk package.
Package: | Rcmdr |
Version: | 2.9-5 |
Date: | 2024-10-23 |
Depends: | R (>= 3.5.0), grDevices, graphics, methods, stats, utils, splines, RcmdrMisc (>= 2.9-1), car (>= 3.1-0), effects (>= 4.0-3) |
Imports: | tcltk, tcltk2 (>= 1.2-6), abind, relimp (>= 1.0-5), lme4, tools |
Suggests: | aplpack, boot, colorspace, e1071, foreign, grid, Hmisc, knitr, lattice, leaps, lmtest, markdown, MASS, mgcv, multcomp (>= 0.991-2), nlme, nnet, nortest, readxl, rgl (>= 0.110.2), rmarkdown (>= 0.9.5), sem (>= 2.1-1) |
ByteCompile: | yes |
License: | GPL (>= 2) |
URL: | https://github.com/RCmdr-Project/rcmdr, https://www.r-project.org, https://www.john-fox.ca/RCommander/index.html |
The R Commander comes with translations from English into several other languages. I am grateful to the following individuals and groups for preparing these translations: Basque, Jose Ramon Rueda; Brazilian Portuguese, Adriano Azevedo-Filho and Marilia Sa Carvalho; Catalan, Manel Salamero; Chinese, Tsungwu Ho, Frank C. S. Liu, and Cheng-shun Lee; Chinese (Simplified), Shulin Yang; French, Philippe Grosjean and Milan Bouchet-Valat; Galician, Anton Meixome; German: Friedrich Leisch and Gerhard Schoen; Greek: Vasileios Dimitropoulos, Anastasios Vikatos, and Andreas Vikatos; Indonesian, I Made Tirta; Italian, Stefano Calza; Japanese, Takaharu Araki; Korean, Chel Hee Lee, Dae-Heung Jang, and Shin Jong-Hwa; Polish, Lukasz Daniel; Romanian, Adrian Dusa; Russian, Alexey Shipunov; Slovenian, Jaro Lajovic and Matjaz Jeran; Spanish, Spanish R-UCA Project, http://knuth.uca.es/R.
John Fox <[email protected]>, Manuel Munoz Marquez, and Milan Bouchet-Valat, with contributions from Liviu Andronic, Michael Ash, Theophilius Boye, Stefano Calza, Andy Chang, Vilmantas Gegzna, Philippe Grosjean, Richard Heiberger, G. Jay Kerns, Renaud Lancelot, Matthieu Lesnoff, Uwe Ligges, Samir Messad, Martin Maechler, Robert Muenchen, Duncan Murdoch, Erich Neuwirth, Dan Putler, Brian Ripley, Miroslav Ristic, Peter Wolf, and Kevin Wright.
Maintainer: John Fox <[email protected]>
Fox, J. (2017) Using the R Commander: A Point-and-Click Interface for R. Chapman and Hall/CRC Press.
Fox, J. (2005) The R Commander: A Basic Statistics Graphical User Interface to R. Journal of Statistical Software, 14(9): 1–42.
Install Pandoc and LaTeX to increase the capabilities of the R Commander.
The capabilities of the R Commander can be enhanced by installing additional software. The R Commander will work without this software but some features will not be activated. The following addiitional software can conveniently be installed via the R Commander Tools > Install auxiliary software menu. The resulting dialog box will take you to websites where the software can be downloaded. This menu item will only be displayed if one or more of these software packages are missing.
Here are the details:
Pandoc: The Pandoc documentation-conversion software is used by the R Commander to generate HTML (web), PDF, and Word files from the editable R Markdown document that is created by default during an R Commander interactive session. Pandoc is required by the rmarkdown package, which, along with the knitr package, performs these conversions. In the absence of Pandoc, R Markdown documents in the R Commander are processed by the older markdown package, which is capable only of producing HTML output. Pandoc is available from https://pandoc.org/installing.html.
On Windows systems, Pandoc installs into a non-standard location in your user directory,
typically C:\Users\<your user id>\AppData\Local\Pandoc\
, and then places this subdirectory
on your system path. You may have to reboot for the change to your path to take effect, and I
have found it necessary on two Windows 10 systems to re-run the Pandoc installer, first to uninstall
Pandoc, and then to re-install it before it would be work.
LaTeX: The LaTeX technical-typesetting system is required by the R Commander for PDF output from R Markdown or knitr documents produced during interactive R Commander sessions. In the absence of LaTeX, direct PDF output is unavailable. Complete LaTeX systems are available for the various platforms that support R and the R Commander, including MikTeX from https://miktex.org/download for Windows systems; MacTeX from https://www.tug.org/mactex/ for MacOSX; and various sources (see https://www.latex-project.org/get/) for Linux/Unix systems.
The CFA dialog is used to create and fit a confirmatory factor analysis model via the
cfa
and sem
functions in the sem package.
Select two or more variables for each factor by Control-clicking on their names in the variable-list box. Optionally give the factor a name; this must be a valid R name. Then press the Define factor button.
Continue in this manner until all factors are specified.
Note that if there are not at least two unique variables selected for each factor, the
model will probably be underidentified, causing sem
to fail.
The radio buttons at the top of the dialog may be used to analyze either the correlation matrix or covariance matrix of the observed variables; to specify either correlated or orthogonal factors; and to identify the model either by setting the factor variance to 1 or by setting the first loading for each factor to 1 (establishing a “reference indicator” for the factor). A check box is provided to compute robust standard errors and tests.
John Fox [email protected]
Start the R Commander GUI (graphical user interface)
Commander()
Commander()
Getting Started
For more detailed information about getting started, see Help -> Introduction to the R Commander from the R Commander menus or Fox (2017).
The default R Commander interface consists of (from top to bottom) a menu bar, a toolbar, a code window with script and R Markdown tabs, an output window, and a messages window.
Commands to read, write, transform, and analyze data are entered using the menus in the menu bar at the top of the Commander window. Most menu items lead to dialog boxes requesting further specification. I suggest that you explore the menus to see what is available.
Below the menu bar is a toolbar with (from left to right) an information field displaying the name of the active data set; buttons for editing and displaying the active data set; and an information field showing the active statistical model. There is also a Submit button for re-executing commands in the Script tab. The information fields for the active data set and active model are actually buttons that can be used to select the active data set and model from among, respectively, data frames or suitable model objects in memory.
Almost all commands require an active data set. When the Commander starts, there is no active data set, as indicated in the data set information field. A data set becomes the active data set when it is read into memory from an R package or imported from a text file, SPSS data set, Minitab data set, STATA data set, SAS XPORT data set; or an Excel spreadsheet. In addition, the active data set can be selected from among R data frames resident in memory. You can therefore switch among data sets during a session.
By default, commands are logged to the Script tab (the
initially empty text window immediately below the toolbar), and
commands and output appear in the Output window (the initially empty text window below the
Script tab). Commands that don't require direct user interaction (such as interactive
identification of points on a graph) are also used to create an R Markdown document in
the tab of the same name. When the R Markdown tab is in front, pressing the "Generate HTML report" button
compiles the document to create an html page with input and output, which opens in a web browser. To alter these and
other defaults, see the information below on configuration. Note, for example, that the knitr package
can be used to create a LaTeX document to be compiled to a PDF report, as an alternative to — or in addition to —
an R Markdown document (see the use.knitr
option below).
Some Rcmdr dialogs (those in the Statistics -> Fit models menu) produce linear, generalized linear, or other models. When a model is fit, it becomes the active model, as indicated in the information field in the R Commander toolbar. Items in the Models menu apply to the active model. Initially, there is no active model. If there are several models in memory, you can select the active model from among them.
If command logging in turned on, R commands that are generated from the menus and dialog boxes are entered into the Script and R Markdown tabs in the Commander. You can edit these commands in the normal manner and can also type new commands. You can also type explanatory text in the R Markdown tab. Individual commands in the Script tab can be continued over more than one line, but the several lines of a multi-line command must be submitted simultaneously. (It is not necessary, as in earlier versions of the R Commander, to begin continuation lines with white space.) The contents of the Script and R Markdown tabs can be saved during or at the end of the session, and a saved script or R Markdown document can be loaded into the respective tabs. The contents of the Output window can also be edited or saved to a text file. Finally, editing operations also work in the Messages window.
To re-execute a command or set of commands in the Script tab, select the lines to be executed using the mouse and press the Submit button at the right of the toolbar (or Control-R, for "run", or Control-Tab). If no text is selected, the Submit button (or Control-R or Control-Tab) submits the line containing the text-insertion cursor. Note that an error will be generated if the submitted command or commands are incomplete.
Pressing Control-F brings up a find-text dialog box (which can also be accessed via Edit -> Find) to search for text in the Script tab, R Markdown tab, knitr tab, Output window, or Messages window. Edit functions such as search are performed in the Script tab unless you first click in another tab or window to make it active.
Pressing Control-S will save the Script tab, R Markdown tab, knitr tab, or Output window.
Pressing Control-A selects all of the text in the Script tab, R Markdown tab, knitr tab, Output window, or Messages window.
In addition, the following Control-key combinations work in these tabs and windows: Control-X, cut; Control-C, copy; Control-V, insert; Control-Z or Alt-Backspace, undo; and Control-W, redo.
Under Mac OS X, the command key may be used in place of the Control key, though the latter works as well.
Right-clicking the mouse (clicking button 3 on a three-button mouse, or Control-left-clicking) in the tabs or windows brings up a "context" menu with the Edit-menu items, plus (in the Script, R Markdown, and knitr tabs) a Submit item.
You can open a larger editor window with the document in the Markdown or knitr tab by making the corresponding selection from the Edit menu, the right-click context menu when the cursor is in the tab, or by pressing Control-E when the cursor is in the tab.
When you execute commands from the Commander window, you must ensure that the sequence of commands is logical. For example, it makes no sense to fit a statistical model to a data set that has not been read into memory.
Pressing a letter key (e.g., "a") in a list box will scroll the list box to bring the next entry starting with that letter to the top of the box.
You can cancel an R Commander dialog box by pressing the Esc key.
Most R Commander dialogs remember their state when this is appropriate, and can be restored to pristine state by pressing the Reset button.
Some R Commander dialogs have an Apply button that will execute the command generated by the dialog and then re-open the dialog in its previous state.
Exit from the Commander via the File -> Exit menu or by closing the Commander window.
Customization and Configuration
The preferred way of customizing the R Commander is to write a plug-in package:
see help("Plugins")
.
Alternatively, configuration files reside in the etc
subdirectory of the package,
or in the locations given by the etc
and etcMenus
options (see below).
The Rcmdr menus can be customized by editing the file Rcmdr-menus.txt
.
You can add R code to the package, e.g., for creating additional dialogs, by placing files with
file type .R
in the etc
directory, also editing Rcmdr-menus.txt
to provide
additional menus, sub-menus, or menu-items. Alternatively, you can edit
the source package and recompile it.
To reiterate, however, the preferred procedure is to write an R Commander plug-in package.
A number of functions are provided to assist in writing dialogs, and Rcmdr state information
is stored in a separate environment. See help("Rcmdr.Utilities")
and the manual supplied in the doc
directory of the Rcmdr package for more information.
In addition, several features are controlled by run-time options, set via the options("Rcmdr")
command. These options should be set before the package is loaded. If the options are unset, which is
the usual situation, defaults are used. Specify options as a list of name=value pairs. You can
set none, one, several, or all options. The available options are as follows:
ask.to.exit
if TRUE
(the default), then the user is asked whether he or she wants to exit the Rcmdr;
if this option is set to FALSE
, then the subsequent option is also set to FALSE
.
ask.on.exit
if TRUE
(the default), then the user is asked whether to save
the script file, R Markdown file, and output file when the Rcmdr exits.
attach.data.set
if TRUE
(the default is FALSE
), the active data set is attached to the
search path.
check.packages
if TRUE
(the default), on start-up, the presence of all of the Rcmdr
recommended packages will be checked, and if any are absent, the Rcmdr will offer to install them.
command.text.color
Color for commands in the output window; the default is "red"
.
console.output
If TRUE
, output is directed to the R Console, and the
R Commander output window is not displayed. The default is FALSE
, unless the R Commander is running under RStudio,
in which case the default is TRUE
.
crisp.dialogs
If TRUE
, dialogs should appear on the screen fully drawn, rather than built up
widget by widget. Prior to R 2.6.1, this option only works on the Windows version of R, but should in any event be harmless.
The default is TRUE
. If you encounter stability problems, try setting this option to
FALSE
.
default.contrasts
Serves the same function as the general contrasts
option; the default isc("contr.Treatment", "contr.poly")
. When the Commander exits, the contrasts
option is returned to its pre-existing value. Note that contr.Treatment
is from the car
package.
default.font.family
The default font for GUI elements such as menus and text labels, in the form of a
Tk font family specification, given in a character string.
For example, "Helvetica"
specifies the sans-serif Helvetica font family.
The default is taken from the TkDefaultFont
. Normally a sans-serif font should be used.
default.font.size
The size, in points, of the default font. The default is 10 on non-Windows
system and the size of the system font on Windows.
To set the font size for R input and output, see the log.font.size
option.
The Rcmdr scale.factor
option may also be used to
control font size.
discreteness.theshold
should be a positive integer; if greater than 0
(which is the default), the maximum number of distinct values
for a numeric variable to be considered discrete; if 0
(or smaller), the threshold is taken as the smallest of 100,
twice the squareroot of the number of cases in the active data set (n), and 10 times log10(n).
double.click
Set to TRUE
if you want a double-click of the left mouse
button to press the default button in all dialogs. The default is FALSE
.
editDataset.threshold
If the number of values in the current
data set exceed this value (the default is 10000), then the standard R
data editor is used in preference to the R Commander editDataset
editor.
error.text.color
Color for error messages; the default is "red"
.
etc
Set to the path of the directory containing the Rcmdr
configuration files; defaults to the etc
subdirectory of the installed Rcmdr package.
grab.focus
Set to TRUE
for the current Tk window to
"grab" the focus — that is, to prevent the focus from being changed to another Tk window.
On some systems, grabbing the focus in this manner apparently causes problems. The default is TRUE
.
If you experience focus problems, try setting this option to FALSE
.
help_type
This Rcmdr option takes precedence over the global R help_type
option (see options
and help
), and by default is set to "html"
.
iconify.commander
If TRUE
, the Commander window is minimized on startup;
the default is FALSE
.
length.output.stack
The R Commander maintains a list of output objects, by default including the last
several outputs; the default length of the output stack is 10. popOutput()
“pops” (i.e., returns and removes) the first entry
of the output stack. Note that, as a stack, the queue is LIFO (“last in, first out”).
length.command.stack
The R Commander also maintains a list of commands that is managed similarly; the default length of this stack is also 10.
log.commands
If TRUE
(the default), commands are echoed to the script window;
if FALSE
, the script window is not displayed.
log.font.family
The font family to be used for text in the script window, output window, messages window,
etc., specified as a character vector giving a Tk font family. This should normally be a monospaced font like "Courier"
.
The default is taken from the TkFixedFont
.
log.font.size
The font size, in points, to be used in the script window,
in the output window, messages window,
in recode dialogs, and in compute expressions — that is, where a monospaced font is used. The default is 10.
Alternatively the Rcmdr scale factor
option may also be used to
control font size.
log.height
The height of the script window, in lines. The default is 10. Setting log.height
to 0 has the same effect as setting log.commands
to FALSE
.
log.text.color
Color for text in the script window; the default is "black"
.
log.width
The width of the script and output windows, in characters. The default is 80.
messages.height
The height of the messages window, in lines. The default is 4.
model.case.deletion
if TRUE
(the default is FALSE
), include a text box for case deletion in statistical-model dialog boxes (e.g., for linear models).
minimum.width
The minimum width, in pixels, for the main R Commander windows; the default is 1000
.
minimum.height
The minimum height, in pixels, for the main R Commander windows; the default is 400
.
multiple.select.mode
Affects the way multiple variables are selected in variable-list boxes.
If set to "extended"
(the default), left-clicking on a variable selects it and deselects any other
variables that are selected; Control-left-click toggles
the selection (and may be used to select additional variables); Shift-left-click extends the selection.
This is the standard Windows convention. If set
to "multiple"
, left-clicking toggles the selection of a variable and may be used to select more than one
variable. This is the behaviour in the Rcmdr prior to version 1.9-10.
number.messages
If TRUE
, the default, messages in the messages window are numbered.
open.graphics.devices
If TRUE
(the default is FALSE
), open the system graphics device and (if 3D RGL graphics are used) the RGL graphics device when the R Commander starts.
open.markdown.editor
If TRUE
(the default is FALSE
), open the R Markdown editor when the R Commander starts.
output.height
The height of the output window, in lines. The default is twice the height
of the script window, or 20 if the script window is suppressed. Setting output.height
to 0 has the same effect as setting console.output
to TRUE
.
output.text.color
Color for output in the output window; the default is "blue"
.
placement
Placement of the R Commander window, in pixels; the default is ""
,
which lets the Tk window manager decide where to place the window; for example,
"+20+20"
should put the window near the upper-left corner of the screen, "-20+20"
near the
upper-right corner, though this doesn't appear to work reliably on Windows systems.
plugins
A character vector giving the names of Rcmdr plug-in packages to load when the Commander starts up. Plug-in packages can also be loaded from the Tools -> Load Rcmdr plug-in(s) menu. See Plugins.
prefixes
A four-item character vector to specify the prefixes used when output is directed
to the R console; the default is c("Rcmdr> ", "Rcmdr+ ", "RcmdrMsg: ", "RcmdrMsg+ ")
.
quit.R.on.close
if TRUE
, both the Commander and R are exited when the Commander
window is closed. The default is FALSE
, in which case only the Commander is exited (and can be restarted
by the command Commander()
).
RcmdrEnv.on.path
If TRUE
(the default is FALSE
),
the environment in which R Commander state information
is stored is placed on the search path. Some plug-ins, at least until
they are updated, may require this setting.
retain.messages
If TRUE
(the default), the contents of the message window
are not erased between messages. In any event, a "NOTE" message will not erase a preceding "WARNING" or "ERROR".
retain.selections
If TRUE
(the default), dialogs remember their previous state, where appropriate,
as long as the data set isn't changed; some dialogs, e.g., for probabilities, retain selections even when the data set chanages.
RExcelSupport
If TRUE
(the default is FALSE
), menus and output are handled by Excel.
rmarkdown.output
Values of several options for converting R Markdown to a document file. The default for this option is TRUE
, which corresponds to markdown.output=list(command.sections=TRUE, section.level=3, toc=TRUE, toc_float=TRUE, toc_depth=3, number_sections=FALSE, translate.rmd.headers=TRUE)
. The sub-option command.sections
controls whether most R commands produce sections in the R Markdown document; the sub-option section.level
controls the level of the sections that are created; the sub-option translate.rmd.headers
controls whether the headers are translated from English into another language, if a translation is available; and the other sub-options are standard for rmarkdown
. The toc_float
, toc_depth
, and number_sections
sub-options are only effective if Pandoc is installed.
rmd.output.format
The output file type for R Markdown documents if pandoc is installed; one of "html"
(the default), "pdf"
(requires LaTeX), "docx"
(Word), or "rtf"
(rich text file).
rmd.template
The quoted path to a .Rmd
file to serve as a template for
R code and output. The default is to use a template included with the package.
scale.factor
A scaling factor to be applied to all Tk elements, such as fonts. This works well
only in Windows. The default is NULL
.
scientific.notation
Higher numbers cause ordinary (decimal) notation to be increasingly preferred to scientific notation
for representing very small and very large numbers; correspond to the scipen
option in R: see options
. The default
is 5
, while the standard default in R is 0
(where 0 means that scientific notation is used whenever the resulting printed
representation of a number is smaller in scientific than in standard notation).
showData.threshold
a vector with 2 entries, defaulting to c(20000, 100)
.
If the number of cases in the active data set exceeds the first number (default, 20,000)
or the number of variables exceeds the second number (default, 100),
then View()
rather than showData()
is used to display the data set. The reason for the option is that showData()
is very slow
when the number of cases or variables is large; setting the threshold to c(0, 0)
suppresses the use of showData
altogether.
It's necessary to use showData
however for the view of the active data set to be updated dynamically when, e.g., a variable is added.
show.edit.button
Set to TRUE
(the default) if you want an Edit button in the Commander
window, permitting you to edit the active data set. Windows users may wish to set this option to FALSE
to suppress the Edit button because changing variable names in the data editor can cause R to crash
(though I believe that this problem as been solved).
sort.names
Set to TRUE
(the default) if you want variable names to be sorted
alphabetically in variable lists.
suppress.icon.images
Set to TRUE
to suppress the icon images in dialog OK, Cancel, Reset, and Help buttons;
the default is FALSE
.
suppress.menus
if TRUE
, the Commander menu bar and tool bar are suppressed, allowing another
program (such as Excel) to take over these functions. The default (of course) is FALSE
.
suppress.X11.warnings
On (some?) Linux and Mac OS X systems, multiple X11 warnings are generated by Rcmdr
commands after a graphics-device window has been opened. Set this option to TRUE
(the default when running
interactively under X11) to suppress reporting of these warnings. An undesirable side effect
is that then all warnings and error messages are intercepted by the Rcmdr,
even those for commands entered at the R command prompt. Messages produced by such commands will be printed
in the Commander Messages window after the next Rcmdr-generated command.
Some X11 warnings may be printed when you exit from the Commander.
theme
A ttk theme to control the overall style of the Commander GUI; should be one of the themes returned by tcltk2::tk2theme.list()
.
The default theme varies by operating system, and can be discovered by entering the command tcltk2::tk2theme()
in a fresh R session.
title.color
Color for the titles of some widgets, such as variable-list boxes; can be given as a color name, such as
"blue"
or as an RGB value, such as "#0000FF"
.
The default is the standard color for ttk label frames, unless that is "#000000"
or "black"
, in which case "blue"
is used instead.
tkwait.commander
This option addresses a problem that, to my knowledge, is rare,
and may occur on some non-Windows systems. If the Commander causes R to hang,
then set the tkwait
option to
TRUE
; otherwise set the option to FALSE
or ignore it.
An undesirable side effect of setting
the tkwait.commander
option to TRUE
is that the R session command prompt is suppressed until the Commander exits.
One can still enter commands via the script window, however. In particular, there is no reason to use this option
under Windows, and it should not be used with the Windows R GUI with buffered output when output is directed to the
R console.
tkwait.dialog
If TRUE
(the default is FALSE
), R will
wait until an R Commander dialog is closed. This has the disadvantage of preventing
help pages from being displayed until a dialog is closed in the Mac OS X R.app
and in RStudio. This was also the standard behavior of the R Commander in earlier
versions and is provided for compatibility with previous behavior. If this option is TRUE
,
then the R Commander data editor is disabled in favor of the standard R platform-specific
data editor, and the new-data-set menu item is suppressed.
use.knitr
If TRUE
(the default is FALSE
), a knitr .Rnw
LaTeX document is created in a tab of the main Commander window;
this document can be compiled into .tex
and .pdf
reports via the knit2pdf
function in the knitr package.
use.markdown
If TRUE
(the default is the negation of the use.knitr
argument), an R Markdown document
is created, which can be compiled into an HTML, PDF, Word, or rich text file report.
use.rgl
If TRUE
(the default), the rgl
package will be loaded if it is present in an
accessible library; if FALSE
, the rgl
package will be ignored even if it is available.
The rgl
package can sometimes cause problems when running R under X11.
"valid.classes"
The classes of variables that the R Commander recognizes, in addition to numeric data; other variables
in a data set will be suppressed. The default is "factor", "ordered", "character",
"logical", "POSIXct", "POSIXlt", "Date", "chron", "yearmon", "yearqtr", "zoo", "zooreg", "timeDate", "xts",
"its", "ti", "jul", "timeSeries", "fts", "Period", "hms", "difftime")
.
variable.list.height
the number of items (typically variables) to display in list boxes; longer lists may be viewed by scrolling. The default is 6.
variable.list.width
a two-item vector controlling the width of list boxes, in characters, giving the minimum and maximum
width to display; the default is c(20, Inf)
. If the widest item name falls in this range, then its number of characters
determines the width of the box. Note: This specification works only approximately.
warning.text.color
Color for warning messages; the default is "darkgreen"
.
Some options can also be set via the File -> Options menu, which will restart the Commander after options are set.
If you want always to launch the R Commander when R starts up, you can include the following code in one of R's start-up
files (e.g., in the Rprofile.site
file in R's etc
subdirectory):
local({ old <- getOption("defaultPackages") options(defaultPackages = c(old, "Rcmdr")) })
R Commander options can also be permanently set in the same manner.
For more information about R initialization, see ?Startup
.
The R Commander Script window does not provide a true console to R, and may have certain limitations. I don't recommend using the R Commander for serious programming or for data analysis that relies primarily on scripts — use a programming editor instead. If you encounter any problems with the Script tab, however, I'd appreciate it if you brought them to my attention.
Under Windows, the Rcmdr package can be run under the Rgui in the SDI (single-document
interface) mode, or under rterm.exe
. You might experience problems running the Rcmdr under
ESS
with NTEmacs or XEmacs, or under other R consoles. The R Commander can be run under the Rgui in
MDI (multiple-document interface) mode but it is relatively inconvenient to do so and isn't recommended.
Occasionally, under Windows, after typing some text into a dialog box (e.g., a subsetting expression in the Subset Data Set dialog), buttons in the dialog (e.g., the OK button) will have no effect when they are pressed. Clicking anywhere inside or outside of the dialog box should restore the function of the buttons. As far as I have been able to ascertain, this is a problem with Tcl/Tk for Windows. I have not seen this behavior in some time and the problem may have been solved.
Under Mac OS X Mavericks and later, the R Commander may appear to freeze or hesitate when run under R.app if the R.app window is hidden and "app nap" is turned on. It is recommended that app nap be turned off for R.app, which can be most conveniently done via the R Commander Tools menu. The app nap setting is permanent until changed and so the current setting will apply whether or not the R Commander is used. When R is first installed, app nap will be on for R.app. The tcltk package requires that X Windows is installed under Mac OS X, and as a consequence the Rcmdr package, which depends on tcltk, will not load if X Windows is absent. X Windows for Mac OS X may be obtained from https://www.xquartz.org/.
On startup, the R Commander sets options(na.action=na.exclude)
; this is done so that
observation statistics such as residuals can be properly added to the active data set
when there are missing values. The
option is reset to its pre-existing value when the Commander exits. Some functions may not
work properly when the default na.action
is set to na.exclude
.
This version should be compatiable with the RExcel package, which can use the R Commander menus.
John Fox [email protected]
Fox, J. (2017) Using the R Commander: A Point-and-Click Interface for R. Chapman and Hall/CRC Press.
Fox, J. (2005) The R Commander: A Basic Statistics Graphical User Interface to R. Journal of Statistical Software, 14(9): 1–42.
Fox, J. (2007) Extending the R Commander by "plug in" packages. R News, 7(3): 46–52.
Plugins, Rcmdr.Utilities, knit
, knit2pdf
options(Rcmdr=list(log.font.size=12, default.contrasts=c("contr.Sum", "contr.poly")))
options(Rcmdr=list(log.font.size=12, default.contrasts=c("contr.Sum", "contr.poly")))
Inicia la GUI (Interfaz Gráfica de Usuario) de R Commander
Commander()
Commander()
Empezando
La interfaz por defecto de R Commander consiste en (de arriba a abajo) una barra de menús, una barra de herramientas, una ventana de instrucciones, una ventana de salida y una ventana de mensajes.
Las instrucciones para leer, escribir, transformar y analizar datos se ejecutan usando la barra de menú de la parte superior de la ventana de R Commander. La mayor parte de los items de este menú le guiarán mediante ventanas de diálogo, preguntando más allá de la especificación. Es aconsejable explorar el menú para ver las opciones disponibles.
Bajo la barra de menú se encuentra la barra de herramientas con (de izquierda a derecha) un campo de información que muestra el nombre del conjunto de datos activos, botones para editar y mostrar el conjunto de datos activos y un campo de información mostrando el modelo estadístico activo. Bajo la ventana de instrucciones hay un botón Ejecutar
para realizar las órdenes indicadas en la ventana de instrucciones. Los campos de información para los datos y el modelo activo son botones que pueden ser usados para seleccionar éstos entre, respectivamente, conjuntos de datos o modelos disponibles en memoria.
La mayor parte de las órdenes requiere un conjunto de datos activos. Cuando se ejecuta R Commander no hay conjunto de datos activos, como está indicado en el campo de información del conjunto de datos activos. Un conjunto de datos llega a ser un conjunto de datos activos cuando éste es leído en la memoria desde un paquete R o importado desde un archivo de texto, conjunto de datos SPSS, conjunto de datos Minitab, conjunto de datos STATA, Excel, Access o dBase. Además el conjunto de datos activos puede ser seleccionado desde conjuntos de datos R residentes en memoria. Los datos pueden ser elegidos de entre todos los conjuntos para cada sesión.
Por defecto, las órdenes son registradas en la ventana de instrucciones (la ventana de texto vacía inmediatamente después de la barra de herramientas); las órdenes y las salidas aparecen en la ventana de resultados (la ventana de texto vacía después de la ventana de instrucciones) y el conjunto de datos activos es adjuntado a la ruta de búsqueda. Para alterar éstos y otros parámetros por defecto, puede consultar la información pertinente en configuración.
Algunos diálogos de Rcmdr
(éstos en Estadisticos -> Ajuste de modelos
) generan el modelo lineal, modelo lineal generalizado y otros modelos. Cuando un modelo es ajustado, se convierte en el modelo activo, indicado en el campo de información de la barra de herramientas de R Commander. Los items del menú Modelos
se aplican al modelo activo. Inicialmente, no hay modelo activo. Si hay varios modelos en memoria, puede elegir el modelo activo de entre ellos.
Si el registro de instrucciones está activo, las órdenes de R generadas desde los menús y los cuadros de diálogos, son introducidas en la ventana de instrucciones de R Comander. Se pueden editar estas órdenes de manera normal y se pueden escribir otras nuevas en la ventana de instrucciones. Las órdenes individuales pueden ser continuadas en más de una línea, pero cada línea después de la primera debe ser identada con uno o más espacios o tabuladores. El contenido de la ventana de instrucciones puede ser almacenado durante o al final de la sesión y un conjunto de instrucciones guardado puede ser cargado en la ventana de instrucciones. El contenido de la ventana de resultados puede ser editado o guardado en un archivo de texto.
Para volver a ejecutar una orden o un conjunto de ellas, se seleccionan las líneas que se desean ejecutar usando el ratón y se presiona el botón Ejecutar
, situado a la derecha de la barra de herramientas (o Control-R
, para ejecutarlos). Si no hay texto seleccionado el botón Ejecutar
(o Control-R
) envía el contenido de la línea que contiene el cursor de insercción. Observar que se generará un error si la orden o las órdenes enviadas son incompletas.
Presionando Control-F
se abre un cuadro de diálogo de búsqueda de texto (también es accesible vía Editar -> Buscar
) para buscar el texto en la ventana de instrucciones o la ventana de resultados. Las búsquedas son realizadas en la ventana de instrucciones a menos que primero pulse en la ventana de resultados para activarla.
Presionando Control-S
se guardará el conjunto de instrucciones o la ventana de resultados.
Presionando Control-A
se selecciona todo el texto del conjunto de instrucciones o de la ventana de resultados.
Pulsando el botón derecho del ratón (el tercer botón en un ratón de tres botones) en el conjunto de instrucciones o en la ventana de resultados se abre el menú contextual con los items del menú Editar
, más un item Ejecutar
(en la ventana de instrucciones).
Cuando ejecute órdenes en la ventana de R Commander, debe asegurarse que la sentencia sea lógica. Por ejemplo, no tiene sentido ajustar un modelo estadístico de un conjunto de datos que no ha sido leído en memoria.
Presionando una letra (ej. "a") en un cuadro con una lista se recorrerá ésta hasta la siguiente entrada que comience con esa letra desde el principio del cuadro.
Salir de R Commander se realiza mediante Fichero -> Salir
o cerrando la ventana de R Commander.
Personalización y configuración
Los archivos de configuración están en el subdirectorio etc
de cada paquete o en la localización dada por etc
y en las opciones de etcMenus
(mirar abajo).
Los menús de Rcmdr
pueden ser personalizados editando el archivo Rcmdr-menus.txt
.
Algunas funciones (ej. histograma
) que normalmente no crean salida visible cuando se ejecutan desde la consola sí lo harán - a menos que se evite - cuando se ejecuten desde la ventana de instrucciones de R Commander. Tal salida puede ser suprimida listando los nombres de estas funciones en el archivo log-exceptions.txt
.
Puede añadir código R al paquete, ej., para crear diálogos adicionales, colocando archivos con extensión .R
en el directorio etc
, además puede editar Rcmdr-menus.txt
para proporcionar menús adicionales, submenús o items. Una demostración de esto se proporciona mediante el archivo BoxCox.demo
. Para activar la demo, renombre el archivo a BoxCox.R
y descomente la correspondiente línea del menú en Rcmdr-menus.txt
. De forma alternativa, puede editar el código del paquete y recompilarlo.
Algunas funciones son proporcionadas para ayudar a escribir diálogos y la información del estado de Rcmdr
en un emplazamiento separado. Mirar help("Rcmdr.Utilities")
y el manual suministrado en el directorio doc
del paquete de Rcmdr
para mayor información.
Además, varias características son controladas mediante opciones, en tiempos de ejecución, establecidas por la orden options("Rcmdr")
. Estas opciones deben ser establecidas antes de cargar el paquete. Si las opciones no están establecidas, que es la situación normal, serán usados los parámetros por defecto. Las opciones se especifican como una lista de pares name$=$values
. Puede no establecer, establecer una, varias, o todas las opciones. Las opciones disponibles son las dadas a continuación:
attach.data.set
Si es TRUE
(por defecto
FALSE
), el conjunto de datos activo es fijado como la ruta de
búsqueda.
check.packages
Si es TRUE
(por defecto), al
arranque, la presencia de todos los paquetes recomendados de Rcmdr
serán comprobados y si alguno no está instalado, Rcmdr preguntará si deben instalarse.
command.text.color
El color de las órdenes en la ventana de resultados es, por defecto, "red"
.
console.output
Si es TRUE
la salida será dirigida a
la consola de R y la ventana de salida de R Commander no se
mostrará. Por defecto es FALSE
.
contrasts
Ofrece la misma función que la opción general
contrasts
; por defecto es c("contr.Treatment",
"contr.poly")
. Cuando se sale de Commander la opción
contrasts
vuelve a su valor preexistente. Observe que
contr.Treatment
es del paquete car
.
crisp.dialogs
Si es TRUE
, los diálogos deben
aparecer en la pantalla dibujada completamente, más que acumular
dispositivo a dispositivo. Esta opción debería afectar sólo a
versiones Windows de R, pero debe en cualquier caso ser
inofensivo. Por defecto es TRUE
bajo versiones Windows de R
2.1.1 y superiores y FALSE
si no. Si está trabajando en
Windows y encuentra que se incrementan los problemas de estabilidad,
pruebe establecer esta opción a FALSE
.
default.font
La fuente por defecto, como la especificación de la fuente de X11, dada en cadena de caracteres. Si está especificado, este valor toma precedencia sobre el tamaño de la fuente por defecto (abajo). Esta opción es sólo para sistemas no-Windows.
default.font.size
Tamaño, en puntos, por defecto de la
fuente. Por defecto es 10 para sistemas Windows y 12 para otro
sistemas, salvo especificación de lo contrario (mirar el item
anterior). La fuente por defecto es
"*helvetica-medium-r-normal-*-xx*"
, donde xx
es por
defecto el tamaño de la fuente. Esta opción es sólo para sistemas no-Windows.
double.click
Establecer a TRUE
si quiere que un doble click con el botón izquierdo del ratón sirva para pulsar el botón por defecto en todos los diálogos. Por defecto es FALSE
.
error.text.color
Color de los mensajes de error; por
defecto es "red"
.
etc
Establece la ruta del directorio que contiene los archivos de configuración de Rcmdr; por defecto el subdirectorio etc
del paquete Rcmdr instalado.
grab.focus
Establecer a TRUE
para "mantener" el enfoque en la ventana actual de Tk, esto es, para prevenir que el enfoque sea cambiado a otra ventana Tk. En algunos sistemas, mantener el enfoque de esta forma, puede causar problemas. Por defecto es TRUE. Si experimenta problemas de enfoque, intente establecer esta opción a FALSE
.
load.at.startup
Vector de caracteres de nombres de los paquetes que deben ser cargados cuando el paquete Rcmdr
es cargado; por defecto se carga sólo el paquete car
. Otros paquetes requeridos serán cargados cuando se necesiten. Si esto está disponible, el paquete car
será cargado cuando Commander se inicie en cualquier caso.
log.commands
Si es TRUE
(por defecto), los comandos son repetidos en la ventana de instrucciones; si es FALSE
, la ventana de instrucciones no se muestra.
log.font.size
Tamaño de la fuente, en puntos, que es usado en la ventana de instrucciones, en la ventana de resultados, en diálogos recodificados y en expresiones de cáculo, esto es, donde es usada una fuente monoespacio. Por defecto es 10 para sistemas Windows y 12 para otros sistemas.
log.height
La altura de la ventana de instrucciones, en líneas. Por defecto es 10. Estableciendo log.height a 0 tiene el mismo efecto que establer log.commands
a FALSE
.
log.text.color
Color del texto de la ventana de instrucciones; por defecto es "black"
.
log.width
La anchura de la ventana de instrucciones y la de salida, en caracteres. Por defecto es 80.
multiple.select.mode
Afecta a la forma en la que múltiples variables son seleccionadas en una caja de listas de variables. Si se establece a "extended"
(por defecto), el botón izquierdo en una variable selecciona ésta y deselecciona cualquier otra variable que estuviera seleccionada; Control+botón izquierdo
acciona la selección (y puede ser usado para seleccionar variables adicionales); Mayúsculas+botón izquierdo
extiende la selección. éste es el convenio estándar de Windows. Si lo establece a "multiple"
, el botón izquierdo acciona la selección de una variable y puede ser usado para seleccionar más de una variable. éste es el comportamiento de Rcmdr antes de la versión 1.9-10.
output.height
Altura de la ventana de resultados, en líneas. Por defecto es dos veces la altura de la ventana de instrucciones o 20 si la ventana de instrucciones es suprimida. Establecer output.height
a 0 tiene el mismo efecto que console.output
a TRUE
.
output.text.color
Color de la salida en la ventana de resultados, por defecto es "blue"
.
placement
Emplazamiento de la ventana de R Commander, en píxeles; por defecto es "$-40+20$"
, lo que pone la ventana cerca de la esquina superior derecha de la pantalla.
plugins
Vector de caracteres con los nombres de paquetes de plugins de Rcmdr a cargar cuando Commander arranque. Los paquetes plugins también pueden ser cargados desde el menú Herramientas -> Cargar paquete(s)
.
suppress.menus
Si es TRUE
, la barra de menús y de herramientas de R Commander son suprimidas, permitiendo que otro programa (como Excel) asuma esas funciones. Por defecto (por supuesto) es FALSE
.
suppress.X11.warnings
En (algunos) sistemas Linux X11 se generan múltiples advertencias por las órdenes de Rcmdr, después de abrir la ventana del dispositivo gráfico.
Establecer esta opción a TRUE
(por defecto cuando arranca interactivamente bajo X11 antes de la versión de R 2.4.0) suprime la aparición de estas advertencias. Un efecto secundario indeseable es que entonces todas las advertencias y mensajes de error son interceptados por Rcmdr, incluso para las instrucciones introducidas en los avisos de R. Los mensajes producidos por tales órdenes serán impresos en la ventana de mensajes de R Commander después de la siguiente orden generada en Rcmdr. Algunas advertencias de X11 puede ser impresas al salir de R Commander. Este problema sólo se aplica a versiones de R anteriores a 2.4.0 y el valor por defecto de la opción es establecido por consiguiente.
retain.messages
Si es TRUE
(por defecto FALSE
), el contenido de la ventana de mensajes no es borrado entre mensajes. En cualquier caso, un mensaje "NOTE" no borrará un anterior "WARNING" o "ERROR".
RExcelSupport
Establecido como TRUE
(por defecto es FALSE
), los menús y salidas son dirigidas a Excel.
scale.factor
Factor de escala aplicado a todos los elementos Tk, como las fuentes. Esto funciona bien sólo en Windows. Por defecto es NULL
.
showData.threshold
Si el número de variables en el conjunto de datos activos excede este valor (por defecto, 100), entonces edit()
, más que showData()
, es utilizado para exhibir el conjunto de datos. Un inconveniente es que el control no se devuelve a Commander hasta que la ventana de edición sea cerrada.
La razón de esta opción es que showData()
es muy lento cuando el número de variables es grande; fijando el umbral a 0 suprime el uso en conjunto de showData.
show.edit.button
Fijar a TRUE
(por defecto) si quiere un botón Editar
en la ventana de Commander, que permita editar el conjunto activo de datos. Los usuarios de Windows pueden desear establecer esta opción a FALSE
para suprimir el botón Editar
porque cambiando los nombres de las variables en el editor de datos se puede causar que R falle (aunque este problema se cree solucionado).
sort.names
Fijar a TRUE
(por defecto) si se quiere ordenar alfabéticamente el nombre de las variables en una lista de variables.
tkwait
Esta opción trata un problema que, en mi
conocimiento, es raro y puede ocurrir en algunos sistemas no
Windows. Si R Commander causa que se cuelgue R, entonces establezca la
opción tkwait
a TRUE
; o conserve la opción en
FALSE
e ignórelo. Un indeseable efecto secundario de establecer
la opción tkwait
a TRUE
es que el aviso de órdenes de la
sesión de R es suprimido hasta salir de R Commander.
Uno sin embargo todavía puede introducir órdenes por la ventana de instrucciones. En
particular, no hay razón para usar esta opción bajo Windows y no se
debería usar con la GUI de R en Windows con salida protegida cuando la
salida esté dirigida a
la consola de R.
use.rgl
Si es TRUE
(por defecto), el paquete rgl
será cargado si está presente en una librería accesible, si es FALSE
, el paquete rgl
será ignorado aunque esté disponible. El paquete rgl puede a veces causar problemas cuando se arranca R bajo X11.
warning.text.color
Color de los mensajes de advertencia;
por defecto es "darkgreen"
.
Muchas opciones pueden también ser establecidas mediante el menú Archivo -> Opciones, que reiniciará R Commander después de que las opciones sean establecidas.
Si quiere lanzar R Commander cuando inicie R, puede incluir la siguiente
instrucción en uno de los ficheros de inicio de R (por ejemplo, en el
fichero Rprofile.site
de la carpeta etc
de R):
local({ old <- getOption("defaultPackages") options(defaultPackages = c(old, "Rcmdr")) })
Las opciones de R Commander puede ser establecidas de forma permanente
de la misma forma. Para más información sobre el inicio de R, véase ?Startup
.
La ventana de instrucciones de R Commander no proporciona una verdadera consola a R y tiene ciertas limitaciones. No se recomienda usar R Commander para la programación sería o el análisis de datos que confíe primordialmente en instrucciones - usar un editor de programación en su lugar.
Por ejemplo, para declaraciones de composiciones de R incluidas entre llaves "\{ \}"
, incluyendo definición de funciones, no serían analizadas ni ejecutadas correctamente, aunque si las líneas después de las primeras que estén identandas. Puede ejecutar declaraciones de composiciones desde la ventana de instrucciones separando los comandos dentro de las llaves por puntos y comas.
Ocasionalmente, bajo Windows, después de teclear algún texto en un cuadro de diálogo (ej. subconjunto de expresiones en el diálogo de subconjunto de conjunto de datos), algunos botones en el diálogo (ej. el botón Aceptar) pueden no tener efecto cuando sean presionados. Pulsando en cualquier sitio, dentro o fuera del cuadro de diálogo, debería restaurarse las funciones de los botones. Por lo que se ha podido comprobar, éste es un problema con Tcl/Tk de Windows.
Bajo Windows, el paquete Rcmdr
puede también funcionar bajo de Rgui en modo SDI (interfaz de único documento) o bajo rterm.exe
; puede ser que experimente problemas ejecutando Rcmdr
bajo ESS
con NTEmacs o XEmacs.
John Fox [email protected] (de la versión inglesa)
Manuel González (traductor) [email protected]
Manuel Muñoz Márquez (traductor–revisor) [email protected]
Véase http://knuth.uca.es/R/doku.php?id=equipotraduccion
options(Rcmdr=list(log.font.size=12, contrasts=c("contr.Sum", "contr.poly")))
options(Rcmdr=list(log.font.size=12, contrasts=c("contr.Sum", "contr.poly")))
The compute dialog is used to compute new variables.
The name of the new variable must be a valid R object name (consisting only of upper and lower-case letters, numerals, and periods, and not starting with a numeral).
Enter an R expression in the box at the right. The expression is evaluated using the active data set. You can double-click in the variable-list box to enter variable names in the expression. The expression must evaluate to a valid variable, which is added to the active data set.
John Fox [email protected]
Allows the user to enter a new dataset, modify data values in an existing dataset, add rows or columns to the dataset, or delete rows or columns.
editDataset(data, dsname, ...) ## S3 method for class 'data.frame' editDataset(data, dsname, ...) ## S3 method for class 'character' editDataset(data, dsname, ...) ## S3 method for class 'NULL' editDataset(data, dsname, ...)
editDataset(data, dsname, ...) ## S3 method for class 'data.frame' editDataset(data, dsname, ...) ## S3 method for class 'character' editDataset(data, dsname, ...) ## S3 method for class 'NULL' editDataset(data, dsname, ...)
data |
an R data frame to edit; this argument is optional, and if absent an empty data frame is created, into which the user can enter data. |
dsname |
the quoted name of the data set, into which the edited data frame
will be placed in the global environment.
If absent and an existing data frame is edited, the modified
version will replace the original version; if absent and a new data set is
created, it will be given the name |
... |
not used by the |
editDataset
is a straightforward spreadsheet-like data editor, suitable for
editing data frames that are not too large (say smaller than about 10,000 values). It is defined as a generic function with a data.frame
method to allow for objects with unique properties that inherit from the data.frame
class. The character
and NULL
methods permit editing an initially empty data set.
Use the mouse and the arrow keys to navigate the cells of the data table, including the row and column names.
Columns consisting only of numbers will produce numeric variables in
the data frame constructed by editDataset
; columns with any non-numeric
values will produce factors or (if they contain only the values
TRUE
and FALSE
) logical variables.
When entering values with embedded blanks, it is permissible but not necessary to enclose the values in quotes
(e.g, "some PS"
or 'less than HS'
).
Clicking in a cell and typing a new value replaces the previous value.
Row and column names can be modified in the same manner.
Double-clicking in a cell deletes the previous value and replaces it with
NA
.
Enlarge the data set by pressing the Add row or Add column button
at the top of the data editor; the new row or column will initially be filled with
NA
s and will have an auto-generated row or column name. Pressing the Enter
or Return key will also add a row; pressing the Tab key will also add
a column.
Right-clicking (or Control-clicking, or under Mac OS X Command-clicking) brings up a context menu, permitting several operations on cells, rows, and columns, including deleting the current row or column.
Similarly, several actions are available via the Edit menu.
The key-combinations Control-x, and Control-v, may also be used respectively to cut, copy, and paste cell values. (Under Mac OS X, Command-x, Command-c, and Command-v also work.)
Pressing the OK button or selecting Exit and save from the File menu exits the data edtior and saves the edited data set to the global environment. Pressing the Cancel button or selecting Cancel from the File menu exits the editor discarding the edited data set.
This function does not return a useful value, but has the side effect of modifying or creating a data set in the global environment.
editDataset
is limited to editing data frames that are composed only of
numeric, factor, and logical columns.
John Fox [email protected]
edit.data.frame
, for the standard R data editor.
if (interactive()) editDataset()
if (interactive()) editDataset()
This dialog is used to specify a generalized linear model to be fit by the
glm
function.
The left model-formula box specifies the response variable to be used in the model; it
may be a variable name or an expression evaluating to the response variable, such as
working == "Fulltime"
.
The right model-formula box specifies the right-hand (i.e., predictor) side of the model.
See glm
for details.
You can type directly in the model formula boxes. Alternatively, double-clicking the left mouse
button on a variable in the variable-list transfers it to the left-hand side of the model (if
it is empty or selected) or to the right-hand side.
Factors are indicated in the variable list; all other variables are numeric.
You can also enter operators and parentheses using the buttons
above the formula. If you select several variables in the variable-list box, clicking on the
+
, *
, or :
button will enter them into the model formula.
Double-click the left mouse button to select a family in the "Family" box and the corresponding
permissible link functions appear in the "Link function" box to the right. Initially, the
canonical link for the family is selected. See family
for details.
Specifying a subset expression allows you to fit the model to a subset of observations
in the active data set. For example, assuming that gender
is a variable in the
active data set, entering gender == "Male"
would restrict the model to males.
The weights box allows you to select a variable specifying prior weights from the drop-down list. Weights giving numbers of trials may be used, for example, to fit a binomial GLM; in this case, the response variable should give the proportion of "successes" for each binomial observation. Click in the weights combo box to see a list of numeric variables in the current data set; type a letter in the box to move the selection cursor to the next variable beginning with that letter.
There is an optional case-deletion box, whose presence is controlled by the model.case.deletion
R Commander option (see Commander
). Typing the row numbers (e,g., 6 16
) or row names (e.g., minister conductor
) of cases to be deleted removes these cases from the fitted linear model. Row names with embedded blanks must be quoted (e.g., "railroad engineer"
), in which case all row names specified should be quoted (e.g., "railroad engineer" "minister"
). You cannot specify both a subset expression and case deletion.
If the active model is a generalized linear model, and the active data set has not changed, then the initial values of the left-hand-side, right-hand-side, family, link, weights, and subset fields are retained from the active model.
John Fox [email protected]
This dialog is used to specify a hierarchical cluster analysis
solution using hclust
, with the distance matrix calculated using
dist
.
Enter a name for the hierarchical clustering solution to be created if you want to retain more than one solution. The solution name must be a valid R object name (consisting only of upper- and lower-case letters, numerials, and periods, and not starting with a number).
Select the variables to be included in the solution using the variable selection box on the left side of the dialog box. A non-contiguous set of variables can be selected by pressing your control key (ctrl) while selecting variables.
Specifying a subset expression (the field below the variable selection box)
allows you to obtain a clustering solution for a subset of observations in
the active data set. For example, assuming that gener is a variable in the
active data set, entering gender == "Male"
would restrict the
solution to males.
Select a clustering method and a distance measure if you are working with raw data. There is often a relationship between the selection of these two items. For example, squared-euclidian distance is appropriate for Ward's method of cluster analysis. If your data is a distance matrix, then select "No Transformation" as the distance measure.
The "Plot Dendrogram" option results in the dendrogram of the solution
being display by using the plot
function.
Dan Putler
This dialog is used to specify a linear model to be fit by the lm
function.
The left model-formula box specifies the response variable to be used in the model; it
may be a variable name or an expression evaluating to the response variable, such as
log(income)
.
The right model-formula box specifies the right-hand (i.e., predictor) side of the model.
See lm
for details.
You can type directly in the model formula boxes. Alternatively, double-clicking the left mouse
button on a variable in the variable-list transfers it to the left-hand side of the model (if
it is empty or selected) or to the right-hand side. You can also enter operators and parentheses using the buttons
above the formula. If you select several variables in the variable-list box, clicking on the
+
, *
, or :
button will enter them into the model formula.
Specifying a subset expression allows you to fit the model to a subset of observations
in the active data set. For example, assuming that gender
is a variable in the
active data set, entering gender == "Male"
would restrict the model to males.
The weights box allows you to perform weight-least-squares (WLS) regression; select a weight variable from the drop-down list. Click in the weights combo box to see a list of numeric variables in the current data set; type a letter in the box to move the selection cursor to the next variable beginning with that letter.
There is an optional case-deletion box, whose presence is controlled by the model.case.deletion
R Commander option (see Commander
). Typing the row numbers (e,g., 6 16
) or row names (e.g., minister conductor
) of cases to be deleted removes these cases from the fitted linear model. Row names with embedded blanks must be quoted (e.g., "railroad engineer"
), in which case all row names specified should be quoted (e.g., "railroad engineer" "minister"
). You cannot specify both a subset expression and case deletion.
If the active model is a linear model and the active data set has not changed, then the initial values of the left-hand-side, right-hand-side, weights, and subset fields are retained from the previous model.
John Fox [email protected]
Plug-ins are R packages that extend the R Command interface.
An R Commander plug-in is an ordinary R package that (1) provides extensions to
the R Commander menus is a file named menus.txt
located in the package's
etc
directory; (2) provides call-back functions required by these menus;
and (3) in an RcmdrModels:
field in the
package's DESCRIPTION
file, augments the list of model objects
recognized by the R Commander. The menus provided by a plug-in package are
merged with the standard Commander menus. It is also possible to remove menus
and menu items from the standard Commander menu file or from the files of
plug-ins installed before the current one.
Plug-in packages given in the R Commander plugins
option
(see Commander
) are automatically loaded when the Commander
starts up. Plug-in packages may also be loaded via the Commander
Tools -> Load Rcmdr plug-in(s) menu; a restart of the Commander is
required to install the new menus. Finally, loading a plug-in package when
the Rcmdr is not loaded will load the Rcmdr and activate the
plug-in.
An illustrative R Commander plug-in package, RcmdrPlugin.TeachingDemos, is available on CRAN.
A variety of utility functions is available to support R Commander plug-in packages;
see Rcmdr.Utilities
.
For more details, see Fox, Writing R Commander Plug-in Packages at https://www.john-fox.ca/RCommander/plug-ins.html.
For technical reasons involving Rcmdr plug-in packages, the Rcmdr package re-exports a number of functions from the tcltk package that are used by Rcmdr utilities.
It also re-exports the showdata
function from the relimp package, which is uses internally to display data sets.
John Fox [email protected]
These functions support writing additions to the Rcmdr package, preferably by writing
an Rcmdr plug-in package. Although it is not recommended, additional R code
can also be placed in files with file type .R
in the etc
subdirectory
of the Rcmdr package. In this case, you can add menus, submenus, and menu items by editing the file
Rcmdr-menus.txt
in the same directory.
activateMenus() activeDataSet(dsname, flushModel=TRUE, flushDialogMemory=TRUE) ActiveDataSet(name) activeDataSetP() activeModel(model) ActiveModel(name) activeModelP() anovaP() beginRmdBlock() beginRnwBlock() Character(names) characterP(n=1) checkActiveDataSet() checkActiveModel() checkBoxes(window=top, frame=stop("frame not supplied"), boxes=stop("boxes not supplied"), initialValues=NULL, labels=stop("labels not supplied"), title=NULL, ttk=FALSE, columns=1) # macro checkClass(object, class, message=NULL) # macro checkFactors(n=1) checkMethod(generic, object, message=NULL, default=FALSE, strict=FALSE, reportError=TRUE) # macro checkNumeric(n=1) checkReplace(name, type=gettextRcmdr("Variable")) checkTwoLevelFactors(n=1) checkVariables(n=1) closeCommander(ask=TRUE, ask.save=ask) closeDialog(window, release=TRUE) # macro Coef(object, ...) Predictors(type=c("all", "numeric", "factor")) PredictorsP(n=1, type=c("all", "numeric", "factor")) CommanderWindow() dataSetsP(n=1) defmacro(..., expr) dialogSuffix(window=top, onOK=onOK, onCancel=onCancel, rows, columns, focus=top, bindReturn=TRUE, preventGrabFocus=FALSE, preventDoubleClick=FALSE, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), tab.names=c("Data", "Options"), grid.buttons=FALSE, resizable=FALSE, force.wait=FALSE) # macro DiscreteNumeric(names) discreteNumericP(n=1) doItAndPrint(command, log=TRUE, rmd=log) EffectP() endRmdBlock() endRnwBlock() enterMarkdown(command) enterKnitr(command) errorCondition(window=top, recall=NULL, message, model=FALSE) # macro exists.method(generic, object, default=TRUE, strict=FALSE) Factors(names) factorsP(n=1) formulaFields(model, hasLhs=TRUE, glm=FALSE) flushDialogMemory(what) gassign(x, value) getCases(cases, remove=TRUE) getDialog(dialog, defaults=NULL) ## S3 method for class 'combobox' getFrame(object) ## S3 method for class 'listbox' getFrame(object) ## S3 method for class 'combobox' getSelection(object) ## S3 method for class 'listbox' getSelection(object) getRcmdr(x, mode="any", fail=TRUE) gettextRcmdr(...) glmP() GrabFocus(value) groupsBox(recall=NULL, label=gettextRcmdr("Plot by:"), initialLabel=gettextRcmdr("Plot by groups"), errorText=gettextRcmdr("There are no factors in the active data set."), variables=Factors(), plotLinesByGroup=FALSE, positionLegend=FALSE, plotLinesByGroupsText=gettextRcmdr("Plot lines by group"), initialGroup=NULL, initialLinesByGroup=1, window=top) # macro groupsLabel(frame=top, groupsBox=groupsBox, columnspan=1, initialText=NULL, ratio=FALSE) # macro hclustSolutionsP() initializeDialog(window=top, title="", offset=10, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), suppress.window.resize.buttons=TRUE) # macro insertRmdSection(text) is.valid.name(x) is.valid.number(string) is.SciViews() justDoIt(command) knitrP() Library(package, pos=length(search()), rmd=TRUE) listAllModels(envir=.GlobalEnv, ...) listAOVModels(envir=.GlobalEnv, ...) listCharacter(dataSet=ActiveDataSet()) listDataSets(envir=.GlobalEnv, ...) listDiscreteNumeric(dataSet=ActiveDataSet()) listFactors(dataSet=ActiveDataSet()) listGeneralizedLinearModels(envir=.GlobalEnv, ...) listLinearModels(envir=.GlobalEnv, ...) listMultinomialLogitModels(envir=.GlobalEnv, ...) listNumeric(dataSet=ActiveDataSet()) listPlugins(loaded=FALSE) listProportionalOddsModels(envir=.GlobalEnv, ...) listTwoLevelFactors(dataSet=ActiveDataSet()) listVariables(dataSet=ActiveDataSet()) lmP() logger(command, rmd=TRUE) logLikP() LogWindow() MacOSXP(release) manualTranslationP() MarkdownP() mavericksP() Message(message, type=c("note", "error", "warning")) MessagesWindow() modelCapability(capability) modelFormula(frame=top, hasLhs=TRUE, rhsExtras=NULL, formulaLabel=gettextRcmdr("Model Formula"), showBar=FALSE) # macro modelsP(n=1) multinomP() Numeric(names) numericP(n=1) OKCancelHelp(window=top, helpSubject=NULL, model=FALSE, reset=NULL, apply=NULL, helpPackage=NULL) # macro OutputWindow() packageAvailable(name) polrP() popCommand(keep=FALSE) popOutput(keep=FALSE) putDialog(dialog, values=NULL, resettable=TRUE) putRcmdr(x, value) radioButtons(window=top, name=stop("name not supplied"), buttons=stop("buttons not supplied"), values=NULL, initialValue=..values[1], labels=stop("labels not supplied"), title="", title.color=getRcmdr("title.color"), right.buttons=FALSE, command=function(){}, columns=1) # macro RappP() RcmdrEditor(buffer, title="R Commander Editor", ok, help=NULL, file.menu=NULL, edit.menu=NULL, context.menu=NULL, toolbar.buttons=NULL) RcmdrTclSet(name, value) RcmdrTkmessageBox(message, icon=c("info", "question", "warning", "error"), type=c("okcancel", "yesno", "ok"), default, title="") removeLastRmdBlock() removeLastRnwBlock() removeNullRmdBlocks() removeNullRnwBlocks() removeStrayRmdBlocks() removeStrayRnwBlocks() RExcelSupported() rglLoaded() RmdWindow() RnwWindow() setBusyCursor() setIdleCursor() sortVarNames(x) subOKCancelHelp(window=subdialog, helpSubject=NULL) # macro subsetBox(window = top, subset.expression = NULL, model = FALSE) # macro suppressMarkdown(command) tclvalue(x) titleLabel(...) tkfocus(...) tkspinbox(parent, ...) trim.blanks(text) TwoLevelFactors(names) twoLevelFactorsP(n=1) UpdateModelNumber(increment=1) variableComboBox(parentWindow, variableList=Variables(), export="FALSE", state="readonly", initialSelection=gettextRcmdr(nullSelection), title="", nullSelection="<no variable selected>", adjustWidth = FALSE) variableListBox(parentWindow, variableList=Variables(), bg="white", selectmode="single", export="FALSE", initialSelection=NULL, listHeight=getRcmdr("variable.list.height"), title) Variables(names) varPosn(variables, type=c("all", "factor", "numeric", "nonfactor", "twoLevelFactor"), vars=NULL) WindowsP() X11P() # the following function is exported for technical reasons, # but are not meant to be called directly commanderPosition()
activateMenus() activeDataSet(dsname, flushModel=TRUE, flushDialogMemory=TRUE) ActiveDataSet(name) activeDataSetP() activeModel(model) ActiveModel(name) activeModelP() anovaP() beginRmdBlock() beginRnwBlock() Character(names) characterP(n=1) checkActiveDataSet() checkActiveModel() checkBoxes(window=top, frame=stop("frame not supplied"), boxes=stop("boxes not supplied"), initialValues=NULL, labels=stop("labels not supplied"), title=NULL, ttk=FALSE, columns=1) # macro checkClass(object, class, message=NULL) # macro checkFactors(n=1) checkMethod(generic, object, message=NULL, default=FALSE, strict=FALSE, reportError=TRUE) # macro checkNumeric(n=1) checkReplace(name, type=gettextRcmdr("Variable")) checkTwoLevelFactors(n=1) checkVariables(n=1) closeCommander(ask=TRUE, ask.save=ask) closeDialog(window, release=TRUE) # macro Coef(object, ...) Predictors(type=c("all", "numeric", "factor")) PredictorsP(n=1, type=c("all", "numeric", "factor")) CommanderWindow() dataSetsP(n=1) defmacro(..., expr) dialogSuffix(window=top, onOK=onOK, onCancel=onCancel, rows, columns, focus=top, bindReturn=TRUE, preventGrabFocus=FALSE, preventDoubleClick=FALSE, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), tab.names=c("Data", "Options"), grid.buttons=FALSE, resizable=FALSE, force.wait=FALSE) # macro DiscreteNumeric(names) discreteNumericP(n=1) doItAndPrint(command, log=TRUE, rmd=log) EffectP() endRmdBlock() endRnwBlock() enterMarkdown(command) enterKnitr(command) errorCondition(window=top, recall=NULL, message, model=FALSE) # macro exists.method(generic, object, default=TRUE, strict=FALSE) Factors(names) factorsP(n=1) formulaFields(model, hasLhs=TRUE, glm=FALSE) flushDialogMemory(what) gassign(x, value) getCases(cases, remove=TRUE) getDialog(dialog, defaults=NULL) ## S3 method for class 'combobox' getFrame(object) ## S3 method for class 'listbox' getFrame(object) ## S3 method for class 'combobox' getSelection(object) ## S3 method for class 'listbox' getSelection(object) getRcmdr(x, mode="any", fail=TRUE) gettextRcmdr(...) glmP() GrabFocus(value) groupsBox(recall=NULL, label=gettextRcmdr("Plot by:"), initialLabel=gettextRcmdr("Plot by groups"), errorText=gettextRcmdr("There are no factors in the active data set."), variables=Factors(), plotLinesByGroup=FALSE, positionLegend=FALSE, plotLinesByGroupsText=gettextRcmdr("Plot lines by group"), initialGroup=NULL, initialLinesByGroup=1, window=top) # macro groupsLabel(frame=top, groupsBox=groupsBox, columnspan=1, initialText=NULL, ratio=FALSE) # macro hclustSolutionsP() initializeDialog(window=top, title="", offset=10, preventCrisp, use.tabs=FALSE, notebook=notebook, tabs=c("dataTab", "optionsTab"), suppress.window.resize.buttons=TRUE) # macro insertRmdSection(text) is.valid.name(x) is.valid.number(string) is.SciViews() justDoIt(command) knitrP() Library(package, pos=length(search()), rmd=TRUE) listAllModels(envir=.GlobalEnv, ...) listAOVModels(envir=.GlobalEnv, ...) listCharacter(dataSet=ActiveDataSet()) listDataSets(envir=.GlobalEnv, ...) listDiscreteNumeric(dataSet=ActiveDataSet()) listFactors(dataSet=ActiveDataSet()) listGeneralizedLinearModels(envir=.GlobalEnv, ...) listLinearModels(envir=.GlobalEnv, ...) listMultinomialLogitModels(envir=.GlobalEnv, ...) listNumeric(dataSet=ActiveDataSet()) listPlugins(loaded=FALSE) listProportionalOddsModels(envir=.GlobalEnv, ...) listTwoLevelFactors(dataSet=ActiveDataSet()) listVariables(dataSet=ActiveDataSet()) lmP() logger(command, rmd=TRUE) logLikP() LogWindow() MacOSXP(release) manualTranslationP() MarkdownP() mavericksP() Message(message, type=c("note", "error", "warning")) MessagesWindow() modelCapability(capability) modelFormula(frame=top, hasLhs=TRUE, rhsExtras=NULL, formulaLabel=gettextRcmdr("Model Formula"), showBar=FALSE) # macro modelsP(n=1) multinomP() Numeric(names) numericP(n=1) OKCancelHelp(window=top, helpSubject=NULL, model=FALSE, reset=NULL, apply=NULL, helpPackage=NULL) # macro OutputWindow() packageAvailable(name) polrP() popCommand(keep=FALSE) popOutput(keep=FALSE) putDialog(dialog, values=NULL, resettable=TRUE) putRcmdr(x, value) radioButtons(window=top, name=stop("name not supplied"), buttons=stop("buttons not supplied"), values=NULL, initialValue=..values[1], labels=stop("labels not supplied"), title="", title.color=getRcmdr("title.color"), right.buttons=FALSE, command=function(){}, columns=1) # macro RappP() RcmdrEditor(buffer, title="R Commander Editor", ok, help=NULL, file.menu=NULL, edit.menu=NULL, context.menu=NULL, toolbar.buttons=NULL) RcmdrTclSet(name, value) RcmdrTkmessageBox(message, icon=c("info", "question", "warning", "error"), type=c("okcancel", "yesno", "ok"), default, title="") removeLastRmdBlock() removeLastRnwBlock() removeNullRmdBlocks() removeNullRnwBlocks() removeStrayRmdBlocks() removeStrayRnwBlocks() RExcelSupported() rglLoaded() RmdWindow() RnwWindow() setBusyCursor() setIdleCursor() sortVarNames(x) subOKCancelHelp(window=subdialog, helpSubject=NULL) # macro subsetBox(window = top, subset.expression = NULL, model = FALSE) # macro suppressMarkdown(command) tclvalue(x) titleLabel(...) tkfocus(...) tkspinbox(parent, ...) trim.blanks(text) TwoLevelFactors(names) twoLevelFactorsP(n=1) UpdateModelNumber(increment=1) variableComboBox(parentWindow, variableList=Variables(), export="FALSE", state="readonly", initialSelection=gettextRcmdr(nullSelection), title="", nullSelection="<no variable selected>", adjustWidth = FALSE) variableListBox(parentWindow, variableList=Variables(), bg="white", selectmode="single", export="FALSE", initialSelection=NULL, listHeight=getRcmdr("variable.list.height"), title) Variables(names) varPosn(variables, type=c("all", "factor", "numeric", "nonfactor", "twoLevelFactor"), vars=NULL) WindowsP() X11P() # the following function is exported for technical reasons, # but are not meant to be called directly commanderPosition()
adjustWidth |
adjust width of combo box to accommodate widest entry (default |
ask |
ask for confirmation. |
ask.save |
ask whether to save contents of script and output windows. |
apply |
if non-null (the default is |
bg |
background color. |
bindReturn |
if |
boxes |
vector of quoted names for check boxes, used to generate each box and its associated variable. |
buffer |
a text string, typically representing the contents of a text widget, such as an R Markdown or knitr document. |
buttons |
vector of quoted names for buttons in a set of related radio buttons. |
capability |
character string giving the name of a column in the R Commander model-capabilities table,
including the name of a column added by a plug-in; see |
cases |
a character string of case number or names to be removed or retained, separated by blanks. |
class |
quoted name of class. |
columns |
number of columns into which to arrange check boxes or radio buttons; boxes and buttons are filled by rows; the default is 1 and the allowed values are 1, 2, 3, and 4. |
columnspan |
number of dialog-box columns to be spanned by frame. |
command |
a character string that evaluates to an R command or (in the case of |
context.menu |
|
dataSet , dsname
|
the quoted name of a data frame in memory. |
default |
default button: if not specified, "ok" for "okcancel", "yes" for "yesno", and "ok" for "ok";
or look for a default method; for |
defaults |
a list of named default values for options in a dialog if no previous selections are stored. |
dialog |
the quoted name of a dialog box under which previous selections are stored. |
edit.menu |
|
envir |
the environment to be searched; should generally be left at the default. |
errorText |
error message to print if a suitable factor isn't available. |
export |
export selection? |
expr |
expression constituting the body of the macro; typically a compound expression. |
fail |
if |
file.menu |
|
flushDialogMemory |
remove saved values of dialog options so that |
flushModel |
set (or reset) the active model to NULL? Should normally be |
focus |
Tk window to get the focus. |
force.wait |
call |
formulaLabel |
text label printed above the formula widget. |
frame |
frame or quoted name for frame depending upon the function. |
generic |
quoted name of generic function. |
glm |
|
grid.buttons |
insert call to |
groupsBox |
listbox object for selecting groups variable. |
initialText |
initial text to display in the groups label; if |
hasLhs |
does the model formula have a left-hand side? |
help |
a two element list: the first element, |
helpSubject |
the quoted name of a help subject, to be called as |
helpPackage |
the quoted name of the package in which to look for help; the default, |
icon |
Message-box icon. |
increment |
increment to model number; -1 to set back after error. |
initialGroup |
quoted name of variable to define groups, set as initial selection in Groups variable list;
|
initialLinesByGroup |
if 1, the lines-by-groups check box is initially checked; 0 to uncheck. |
initialLabel |
label for groups button before a selection is made. |
initialSelection |
index of item initially selected, 0-base indexing. |
initialValue |
for a set of related radio buttons. |
initialValues |
for a set of related check boxes. |
keep |
if |
label |
label prefix for groups button after a selection is made. |
labels |
a vector of character strings to label a set of radio buttons or check boxes. |
listHeight |
Maximum number of elements displayed simultaneously in list box. |
loaded |
if |
log |
echo command to the script window, as well as executing it and printing its output. |
message |
error (or other) message. |
mode |
mode of object to retrieve. |
model |
the name of a model, as a character string, or a model object,
or |
n |
number of items to check for. |
name |
quoted name. |
names |
optional names to be stored. |
notebook |
notebook windows for a tabbed dialog (default |
nullSelection |
what user selects in combo box to indicate nothing selected (default |
object |
an object (depends on context). |
offset |
in pixels, from top-left of Commander window. |
ok |
a function called when the OK button is pressed in a |
onOK |
function to execute when the OK button is pressed. |
onCancel |
function to execute when the Cancel button or Esc key is pressed. |
package |
quoted name of package to load. |
parent |
parent Tk window. |
plotLinesByGroup |
include a check box for plotting lines by group? |
plotLinesByGroupsText |
the label for the plot-lines-by-group check box. |
pos |
position on search path at which to load package; default is just before the end of the path. |
positionLegend |
include a check box for a legend? |
preventGrabFocus |
prevent the dialog box from grabbing the focus. |
preventDoubleClick |
prevent double-clicking from pressing the OK button, even when the double.click option is set; necessary for statistical modelling dialogs, which use double-clicking to build the model formula. |
preventCrisp |
this argument is ignored, and is present only for backwards compatibility. |
ratio |
If |
recall |
function to call after error — usually the function that initiates the dialog. |
release |
release the focus if the grab.focus option has been set. In
|
remove |
If |
reportError |
if |
reset |
quoted name of dialog function, to be invoked with all defaults by Reset button. |
resettable |
should dialog state be reset when the data set changes? The default is
|
resizable |
should the dialog be resizable by the user? The default is |
rhsExtras |
show controls for splines and polynomials for a model formula; for backwards
compatibility, defaults to |
right.buttons |
radio button placed to right of button-labels; defaults to |
rmd |
enter the command in the R Markdown tab. |
rows |
numbers of rows of widgets in the dialog box; this is actually no longer used, but
is still present for backwards compatibility. The |
selectmode |
|
showBar |
include a bar (|) button in the model-formula widget (default is |
state |
state of the combobox widget; default |
strict |
if |
string |
a character string, or vector of strings, to be tested whether it can be coerced to a number or numbers;
returns either |
subset.expression |
a quoted expression to subset the data set. |
suppress.window.resize.buttons |
if |
tab.names |
text to print as tab labels (default |
tabs |
quoted names of tabs for a tabbed dialog (default |
text |
a text string. |
title |
Window or dialog-box-element title. |
title.color |
color for title above radio buttons; defaults to |
toolbar.buttons |
|
ttk |
use ttk themed widget for check boxes. |
type |
quoted type of object to check; used to generate check-replace dialog box;
or type of message to print in Message window. For |
use.tabs |
(default |
value |
an object to be stored or assigned. |
values |
vector of quoted values associated with radio buttons or check boxes; for |
variableList |
a vector of variable names. |
variables |
a vector of one or more variable names. |
vars |
a vector of one or more variable names; if given |
what |
optional character vector of one or more dialog names for which the memory is to be flushed; if not specified, all dialog memory will be flushed. |
window , parentWindow
|
a Tk window. |
x |
an R object name, as a character string, or a tcl variable or object, or a vector of variable names to be sorted. |
... |
For |
There are several groups of functions exported by the Rcmdr package and documented briefly here. To see how these functions work, it is simplest to examine the dialog-generating functions in the Rcmdr package. Also see the RcmdrPlugin.survival package for examples.
Executing and logging commands:
The functions doItAndPrint
, justDoIt
, and logger
control the
execution, logging, and printing of commands generated by menus and dialogs.
logger(command)
adds command
to the log/script window and to the output window.
justDoIt(command)
causes command
to be executed. doItAndPrint(command)
does both of these operations, and also prints the
output produced by the command.
The R Commander maintains a list of output objects, by default including the last
10 outputs. popOutput()
“pops” (i.e., returns and removes) the first entry
of the output stack. Note that, as a stack, the queue is LIFO (“last in, first out”). Use popOutput(keep=TRUE)
to access the last output but keep it in the stack.
There is also a stack of commands, which is accessed similarly by popCommand()
.
Ocassionally, it's necessary to assign an object directly in the global environment, and
this can be done with the gassign
function.
Normally commands also generate an R Markdown block.
suppressMarkdown
takes a command in character-string form and adds an attribute to it
that will cause the command not to be entered in the R Markdown tab. This is useful when a
command, such as identify
, requires direct user interaction and won't generate useful Markdown.
enterMarkdown
can be used to enter command blocks directly in the R Markdown tab; this should
rarely be required. The functions beginRmdBlock
, endRmdBlock
, removeNullRmdBlocks
,
removeLastRmdBlock
, and removeStrayRmdBlocks
should normally not be called directly.
The functions enterKnitr
,
beginRnwBlock
, endRnwBlock
, removeNullRnwBlocks
, removeLastRnwBlock
,
and removeStrayRnwBlocks
perform similar functions for Knitr documents.
insertRmdSection
Inserts a Markdown section title immediately above the last R command block, with the specified text as the title. In most instances it's unnecessary to do this directly because most commands automatically generate a Markdown section title.
Checking for errors:
The function is.valid.name
checks whether a character string specifies a valid name
for an R object. The function is.valid.number
checks whether a character string (or vector) can be coerced to a
number (or numbers). The functions checkActiveDataSet
, checkActiveModel
,
checkFactors
, checkNumeric
, checkTwoLevelFactors
, and checkVariables
check for the existence of objects and write an error message to the log if they are absent
(or insufficiently numerous, in the case of different kinds of variables). The function
checkReplace
opens a dialog to query whether an existing object should be replaced.
The function checkMethod
, checks whether a method exists for a particular generic that is
appropriate for a particular object. The function checkClass
checks whether an object is of
a specific class. Both of these functions write error messages to the log if the condition fails.
The function errorCondition
reports an error to the user and (optionally) re-starts
a dialog.
Information:
Several functions return vectors of object names: listAllModels
, listAOVModels
, listCharacter
, listDataSets
, listDiscreteNumeric
, listGeneralizedLinearModels
,
listFactors
, listLinearModels
, listMultinomialLogitModels
,
listNumeric
, listProportionalOddsModels
,
listTwoLevelFactors
,
listVariables
. The functions Predictors
and Coefficients
return information about the active model, or NULL
if there is no active model. The functions activeDataSet
and activeModel
respectively report or set the active
data set and model.
The function packageAvailable
reports whether the named package is
available to be loaded (or has possibly already been loaded).
The function exists.method
checks whether a method exists for a particular generic that is
appropriate for a particular object, and returns TRUE
or FALSE
. The function is.SciViews
always returns FALSE
since the SciViews GUI is no longer supported.
The function modelCapability()
returns TRUE
if there is an active statistical model and if it has the specified capability.
For example, modelCapability("sum")
returns TRUE
if the model-capabilities table indicates that there's an applicable
summary()
method for the active model. Otherwise, FALSE
is returned. If the specified capability doesn't exist, a warning is printed.
Building dialog boxes:
Several functions simplify the process of constructing Tk dialogs:
initializing a dialog box, initializeDialog
, and completing the definition of a dialog box, dialogSuffix
;
a set of check boxes, checkBoxes
;
a set of radio buttons, radioButtons
;
a list box with associated scrollbars and state variable, variableListBox
(and associated methods for the generic functions
getFrame
and getSelection
);
a drop-down "combo" box, variablecomboBox
(with getFrame
and getSelection
methods);
a button and subdialog for selecting a "grouping" variable, groupsBox
;
displaying the currently defined groups in a dialog, groupsLabel
;
a dialog-box structure for entering a model formula, modelFormula
;
a text box for entering a subsetting expression, subsetBox
;
OK, Cancel, and Help buttons for dialogs, OKCancelHelp
, and subdialogs, subOKCancelHelp
.
The functions putDialog
, getDialog
, and varPosn
support dialog-box memory—i.e., retaining
selections across invocations of a dialog. The tkspinbox
function is omitted from the tcltk package and
may be used to create a spinbox widget. The titleLabel
function may be used to format a title label to use
the standard title label font and color.
“Themed” Tk widgets:
Tk 8.5 introduced so-called “themed” widgets, which look better than the traditional Tk widgets.
Several functions, contributed by Brian Ripley, are written to access the new widgets by switching automatically
between the new and old widget sets depending upon the availability of the former: buttonRcmdr
,
to access either ttkbutton
or tkbutton
; labelRcmdr
,
to access either ttklabel
or tklabel
; ttkentry
,
to access either ttkentry
or tkentry
; ttkframe
,
to access either ttkframe
or tkframe
; ttkradiobutton
,
to access either ttkradiobutton
or tkradiobutton
; and ttkscrollbar
,
to access either ttkscrollbar
or tkscrollbar
.
Note that the last four functions mask functions of the same names in the tcltk package.
‘Predicate’ functions:
A number of functions of the form nameP
are ‘predicate’ functions, which return
TRUE
or FALSE
depending upon whether some condition obtains. For example, lmP()
returns TRUE
if there is an active model that is a linear model; and factorsP(2)
returns TRUE
if there are at least two factors in the active data set.
WindowsP()
, MacOSXP()
, and X11P()
return TRUE
if the R Commander is running under Windows, Mac OS X, or
X-Windows, consecutively.
Translating text:
The gettextRcmdr
function simply passes its argument(s) to
gettext
, adding the argument domain="R-Rcmdr"
It is not meant to be used in plug-in packages.
Miscellaneous:
The function trim.blanks
removes spaces from the beginning and end of a character string. The function
installPlugin
installs an Rcmdr plug-in from a ZIP file or directory; this function may be used to
create self-installing plug-ins in the form of packages. The function nobs
returns the number of
observations on which a statistical model is based. The function formulaFields
returns information
about the left-hand side, right-hand side, data, subset, and (for GLMs) family and link, of a model object.
The function sortVarNames
sorts variable names, including those containing numerals, into a more
"natural" order than does the standard sort
function. The function Library
may be used to
load packages; it checks whether a package is already loaded, and if not by default puts it in position
4 on the search path. The function Coef
, with several methods, returns the coefficients of a model
as a vector; the default method just calls coef
.
The function RExcelSupported
is used for the RExcel interface.
The function getCases
takes a character string of case names or numbers as an argument and returns a character string that evaluates to an expression either to delete or retain these cases.
Some of these functions, marked # macro
under Usage, are "macro-like" in their behaviour, in that they
execute in the environment from which they are called. These were defined with an adaptation (used with permission) of
Thomas Lumley's defmacro
function, described in Lumley (2001), and are used in the R Commander to deal with
scoping issues related to Tcl/Tk.
The tkfocus
function is exported for historical reasons.
John Fox [email protected]
T. Lumley (2001) Programmer's niche: Macros in R. R News, 1(3), 11–13.
This is a slightly modified version of the tkpager, changed to use the Rcmdr monospaced font and a white background.
RcmdrPager(file, header, title, delete.file)
RcmdrPager(file, header, title, delete.file)
file |
character vector of file(s) to be displayed. |
header |
for the beginning of each file. |
title |
for window |
delete.file |
delete file(s) on close. |
The recode dialog is normally used to recode numeric variables and factors
into factors, for example by combining values of numeric variables or levels
of factors. It may also be used to produce new numeric variables. The Rcmdr
recode dialog is based on the Recode
function in the
car
package.
The name of each new variable must be a valid R object name (consisting only of upper and lower-case letters, numerals, and periods, and not starting with a numeral).
Enter recode directives in the box near the bottom of the dialog.
Directives are normally entered one per line,
but may also be separated by semicolons. Each directive is of the form input = output
(see the examples below). If an input value satisfies more than one specification, then the
first (from top to bottom, and left to right) applies. If no specification is satisfied, then
the input value is carried over to the result. NA
is allowed on input and output. Factor
levels are enclosed in double-quotes on both input and output.
Several recode specifications are supported:
For example, "missing" = NA
.
For example, 7,8,9 = "high"
.
For example, 7:9 = "high"
. The
special values lo
and hi
may appear in a range. For example,
lo:10=1
. Note that these values are unquoted.
else
everything that does not fit a previous
specification. For example, else=NA
. Note that else
matches all
otherwise unspecified values on input, including NA
.
If all of the output values are numeric, and the "Make new variable a factor" check box is unchecked, then a numeric result is returned.
If several variables are selected for recoding, then each is recoded using the same recode directives. In this case, the name entered in the box labelled "New variable name or prefix for multiple recodes" will be prefixed to the name of each variable being recoded. Setting an empty prefix (i.e., "") will cause the recoded variables to replace the original variables.
As explained, =
is used to separate old from new values, and :
is used to
specify an interval (or range) of numeric values. It is possible to change these operators
to other character strings, such as ->
and ~
(tilde). This may be necessary, for example, if a factor to be
recoded has =
s or :
s in its level (category) names.
Similarly, the dialog generates a call to the Recode
function in the car package, which by
default uses ;
to separate recode specifications. The recode separator can also
be changed, for example, to /
.
John Fox [email protected]
The one-way and two-way repeated-measures ANOVA/ANCOVA dialogs compute analysis of variance and analysis of covariance tables for one or two repeated-measures factors and a between-subjects linear model that can include both factors and covariates.
The data are assumed to be in "wide" format, with repeated measures corresponding to distinct variables in the active data set. If the data are in "long" format, they can be reshaped to wide format via the Data > Active data set > Reshape data set from long to wide format dialog.
The model is specified in the Design tab in two parts:
The within-subjects design is defined by using the drop-down lists to select the variables that correspond to the levels of the within-subjects factor (in the case of one repeated-measures factor) or the combinations of levels of the within-subjects factors (in the case of two repeated-measures factors, organized as a two-way table). The user can also name the levels of the within-subjects factor(s) and the factor or factors themselves.
The between-subjects model is specified exactly as for a linear model (see linearModel
). The model is then fit as a multivariate linear model with the repeated measures as response variables.
In the Options tab, the user can select either multivariate tests (using one of four test statistics) or univariate tests. The tests are performed by the Anova
function in the car package.
The means of the repeated measures can optionally be plotted as a function of within- and between-subjects factors, and the means and standard deviations can be printed, using the repeatedMeasuresPlot
function in the RcmdrMisc package.
John Fox [email protected]
John Fox, Michael Friendly, and Sanford Weisberg, "Hypothesis Tests for Multivariate Linear Models Using the car Package", The R Journal, 5 (1) 39–52, 2013.
linearModel
, Anova
, repeatedMeasuresPlot
There are two dialogs for "reshaping" the active data set: (1) from "long" to "wide" format (using the reshapeL2W
in the RcmdrMisc package); and (2) from "wide" to "long" format (using the reshapeW2L
in the RcmdrMisc package).
These dialogs are designed for handling regularly structured repeated-measures data, where each subject (independent unit of observation) is measured on several occasions or under several different conditions. The occasions may have either a one-dimensional structure (corresponding to the levels of one repeated-measures or within-subjects factor) or a two-dimensional structure (corresponding to two crossed repeated-measures factors).
Data in "wide" format have one row for each subject, with the repeated measures in different columns (variables). Data in "long" format have several rows for each subject, with one column (variable) for the response; the levels of the repeated-measures factor (or combinations of levels for two repeated-measures factors) correspond to distinct rows. The within-subjects factor or factors appear as different columns in the long data, as do the between-subjects factors; the former vary within subjects, while the latter are invariant within subjects; and an ID variable identifies the rows of the data set belonging to each subject.
Data in wide format are suitable for analysis by the R Commander repeated-measures ANOVA/ANCOVA dialogs (see RepeatedMeasuresDialogs
), while data in long format are suitable for analysis by the R Commander mixed-models dialogs.
The Reshape Data Set from Long to Wide Format dialog is largely self-explanatory: The user selects the variable that identifies subjects (i.e., the ID variable); one or two within-subjects factors; the variable or variables that vary by occasion (there is typically just one, the response variable); and any variables in the long data set that are to be excluded from the wide data set.
The Reshape Data Set from Wide to Long Format dialog is more complex. There are three tabs: A tab to specify one repeated-measures factor; a tab to specify two crossed repeated-measures factors; and an options tab. The user specifies either one or two repeated-measures factors, not both. The within-subjects factor or factors are defined by using drop-down lists to select the variables that correspond to the levels of the within-subjects factor (in the case of one repeated-measures factor) or the combinations of levels of the within-subjects factors (in the case of two repeated-measures factors, organized as a two-way table). The user can also name the levels of the within-subjects factor(s) and the factor or factors themselves.
John Fox [email protected]
reshapeL2W
, reshapeW2L
, RepeatedMeasuresDialogs
.
This dialog creates a ‘.Rprofile’ file, by default in the current directory, adding to it the current R Commander options, set, e.g., in the Options dialog. If R is subsequently restarted in this directory, and the Rcmdr package loaded, then the current R Commander options will be applied. The current directory will typically, though not necessarily, be your home directory — for example, for Windows users, your Documents directory.
If a ‘.Rprofile’ file already exists in the current directory, then the R Commander options are added to it at the end of the file, after removing R Commander options previously generated by an earlier invocation of this dialog.
The contents of the ‘.Rprofile’ file may be edited before you save the file.
If you want to start the R Commander automatically when R starts, uncomment (remove the #
s from)
the four lines
# local({ # old <- getOption('defaultPackages') # options(defaultPackages = c(old, 'Rcmdr')) # })
See Startup
for a description of the ‘.Rprofile’
file and the R startup process in general.
See Commander
for a description of the various R Commander options.
saveOptions() # typically through the R Commander Tools menu
saveOptions() # typically through the R Commander Tools menu
John Fox [email protected]
This dialog sets up a call to the scatter3d
function to draw a
three-dimensional scatterplot, and optionally to Identify3d
to label
points interactively with the mouse.
The explanatory variables provide the "horizontal" and "out-of-screen" axes of the scatterplot, the response variable provides the "vertical" axis.
Data points are represented as spheres or points, depending upon the number of observations.
Several regression surfaces can be
plotted: a linear least-squares surface; a full quadratic least-squares surface
with squared and cross-product terms; a "smooth" regression surface — either a
smoothing spline, if no degrees of freedom are specified (in which case the
gam
function selects the df by generalized cross validation),
or a fixed-df regression spline; an additive-regression surface (also fit by gam
),
with either smoothing spline or regression spline components (again selected according
to the specification of degrees of freedom). If only one surface is fit, then residuals are
plotted as red (negative) and green (positive) lines from the surface to the points.
You can specify a factor defining groups by pressing the Plot by groups button. A separate surface or set of surfaces is plotted for each level of the groups factor. These surfaces can be constrained to be parallel.
The completed plot can be manipulated with the mouse: Click, hold, drag the left mouse button to rotate the display; click, hold, and drag the right button (or centre button on a three-button mouse) to zoom in and out.
If the box labelled Identify observations with mouse is checked, you may use the mouse to identify points interactively: Press the right mouse button (or the centre button on a three-button mouse), drag a rectangle around the points to be identified, and release the button. Repeat this procedure for each point or set of "nearby" points to be identified. To exit from point-identification mode, right-click (or centre-click) in an empty region of the plot.
Points may also be identified subsequently by selecting Identify observations with mouse from the R Commander 3D graph menu: As above, click and drag the left mouse button to rotate the display, and click and drag the right (or centre) button to identify points.
John Fox [email protected]
scatter3d
, Identify3d
,
rgl-package
, gam
The R Commander script editor is meant to edit scripts in text widgets, such as the R Commander R Markdown and knitr document tabs.
Saving the document, either via the File menu or pressing the OK button closes the editor and modifies the content of the corresponding R Markdown or knitr tab. Closing the editor without saving, by selecting Cancel from the file menu, pressing the Cancel button or destroying the window discards changes to the document. You may also save your edits to the R Markdown or knitr tab without closing the editor. Compiling the document into a report also saves the current edits.
The editor is a “non-modal” dialog, and so may remain open when you work.
John Fox [email protected]
using R Markdown; using knitr.