Windows binaries at CRAN (bin/windows/Windows-NT/base) have been updated to R-1.1.0. See below for a list of Windows-specific changes. We thanks all the people who checked over pre-test versions. guido masarotto (for the R-core team) Windows-specific changes to R ============================ There is now a GUI preferences editor on the Edit menu in Rgui. A data entry editor is now implemented. This is similar to the Unix one (see ?dataentry: that has itself been considerably enhanced in parallel with the Windows version) with the additional features that double-clicking in a cell turns it into an editable field (entered as a new value by navigating away from the cell), and there is a right-click popup menu. Rterm will now load and save the commands history (as Rgui does). There is a new item on the File menu in Rgui to save the commands history at that point in a session, and a new function savehistory() (only applicable in Rgui and Rterm). Colours in Rconsole can be specified as "Windows" and will be taken from the current Windows colour scheme. (The preferences editor will save the actual colour, though.) The placement of multiple pagers is now `random' (so they will not always appear in the exact centre of the screen) and under the MDI interface the pager rows and columns settings are now used. The postscript() device can now spool plots to a printer. You will need a suitable program to send a file to the spooler (or even to send it to ghostscript to interpret PostScript for the printer): the RedMon suite available from http://www.cs.wisc.edu/~ghost/rjl.html and GhostScript distribution sites should provide suitable tools. See ?postscript for further details. The new xfig() device can be used, and a port of xfig to Windows (with an X server and the Cygwin B20.1 environment and X11R6.4 installed) is at ftp.franken.de:/pub/win32/develop/gnuwin32/cygwin/porters/Humblet_Pierre_A/B20 as xfig-3.2.2-cygwin-b20.1-bin.tar.gz. The R DLL and executables now contain standard Windows version information, which will appear on a Version tab of the Properties dialog box in Explorer. The internal unzip and the installer now use unzip5.41, not unzip5.40, and the DLL interface is incompatible. There are changes in the tools needed to build R (see the file INSTALL) and those recommended for building packages from source (see the file readme.packages). New makefile target zipdata-% to zip up data directories in packages. If there is a file `Makefile.win' in a package src directory this will be used rather than MakePkgs (and the tcltk package provides an example). There is an example package `windlgs' in the package sources distribution that shows how to use graphapp to build dialog boxes to be called from menus, with a worked example of a t test done in three ways. More details of graphapp, including a tutorial and a reference manual, are at http://www.cs.usyd.edu.au/~loki/GraphApp. Optional support for the Tcl/Tk interface. To use this you need to install tcl831.exe from dev.scriptics.com, and you will need the Tcl and Tk dlls in your path. Also, you need to set the environment variable TCL_LIBRARY, to something like c:/Program Files/Tcl/lib/tcl8.3. The demos(tk*) will work on Windows. The R mathematical functions can now be built as a DLL for use with your own C programs. See src/nmath/standalone/README in the sources. Rdconv.bat no longer depends on having a Unix-compatible `cat' program in your path. A long-standing bug in the (vertical) alignment of symbols in plotmath has been corrected. Pagers should no longer flash when launched. Launching a pager or data editor or graphics window now brings it to the front, even from Rterm.exe. (Not for Tcl/Tk windows, unfortunately.) Failure to load a DLL with dyn.load now gives extended error information. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-announce mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-announce-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Again the same type of problem with scoping.
I wrote the simple functions below, 'by' is from Trevor Hastie
modified by F.Harrel and copied from the list.
epi.crosstab <- function(vec1, vec2, vec3=NULL, row.labels =
NULL, col.labels = NULL)
{
if(!is.null(vec3) ){
data <- eval(epi.file,sys.frame(sys.parent()))
by(vec3, xtab(vec1,vec2, row.labels = row.labels, col.labels =
col.labels),data=data)
}
if(is.null(vec3))
xtab(vec1, vec2, row.labels = row.labels, col.labels=
col.labels)
}
by <- function(group, exp, data=data)
{
G <- substitute(group)
exp <- substitute(exp)
G <- factor(eval(G))
for(group in levels(G)) {
eval(exp, envir=c(data[G == group, ]))
}
invisible()
}
xtab <- function(vector1, vector2,row.labels=NULL,col.labels=NULL)
{
table(vector1, vector2)
}
now I have a data.frame that is always called 'epi.file' and is
attached to frame 1, MILK, COFFEE and WATER are variables in
the frame
> epi.crosstab(MILK,COFFEE)
vector2
vector1 N Y
N 44 27
Y 0 4
==Why it does print vector1, vector 2 and not MILK and COFFEE ?
> epi.crosstab(MILK,COFFEE,WATER)
Error in eval(expr, envir, enclos) : Object "vec3" not
found>
Can somebody be so kind as to try to explain which scoping rule
now I did violate and where should I put 'eval' 'substitute' etc
so it
works.
Thanks
.
R. Heberto Ghezzo Ph.D.
Meakins-Christie Labs
McGill University
Montreal - Canada
heberto at meakins.lan.mcgill.ca
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Hi, I wrote the simple program below, I want 'bartlett.test' to print
the
real name of the variables, not the dummys 'y' and 'g', Can
somebody in the list tell me how to pass the names?
.
test <- function(y,g){
# some other stuff
bartlett.test(y,g)
# some other stuff
}
Thanks.
R. Heberto Ghezzo Ph.D.
Meakins-Christie Labs
McGill University
Montreal - Canada
heberto at meakins.lan.mcgill.ca
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._