[R] .ps and .pdf page size differences in FC vs. Debian

Clint Harshaw charshaw at presby.edu
Wed Mar 2 19:10:43 CET 2005


R version: 2.0.0
OS: Fedora Core 1

When I'm using Fedora Core 1, Rplots.ps (and subsequent Rplots.pdf 
created after using ps2pdf Rplots.ps) which are created from the command 
line execution of R are too large to fit on a US Letter sized page when 
printed. The same code will produce a US letter sized page when I'm on 
my Debian Sid box. Both boxes are connected to the same shared printer, 
and both are configured in CUPS to use US Letter paper size. When I 
posted a question to the FC folks, the consensus was there was a 
configuration that needed to be done in R to set the proper page for 
graphics exported.

I've google'd for a solution and came up with a check for a locale 
variable in /etc/sysconfig/i18n to be set to the same as for the lang 
variable. I tried that, and the results were still incorrectly sized.

Three sample graphs of the problem can be seen at this url:

http://penguinsolutions.org/math108/Rexamples/Rplots.pdf

Notice how the graph extends beyond the paper on the right side of each 
page.

The command I am running from the terminal is:

R <baseballsalary.R >output --vanilla -q

I then read about postscript() in the documentation, so I found this 
sequence to be successful:

 > postscript("testthis.ps", paper='letter')
[...run all my same R code here...]
 > dev.off()
 > q()

and then run ps2pdf testthis.ps to create the plots as I intended on US 
letter:

http://penguinsolutions.org/math108/Rexamples/testthis.pdf

Notice that the graphs do not extend beyond the edge of US letter paper 
on these plots.

How do I set (in Fedora Core 1) the paper size for R to US letter, so I 
can run my commands R <baseballsalary.R >output --vanilla -q from the 
terminal?

Thanks very much,
Clint
-- 
Clint Harshaw, PhD   		Email: charshaw at presby.edu
Department of Mathematics	Ph: 864.833.8995
Presbyterian College		Fax: 864.938.3769	
Clinton, SC USA  29325		Harrington-Peachtree Rm 412




More information about the R-help mailing list