Hi everyone, I have been making a fair amount of figures in R recently that I've been touching up with Illustrator and I've found a difference between pdf and ps files and I was wondering if someone could enlighten me about them. While the figures look the same, the ps version tends to have truncated strings. The last character of short strings tends to be on a string of its own, located right beside the rest. This makes it a bit awkward to manipulate, especially if scaling is involved. Is there a reason for this differences? There also seems to be somewhat arbitrary grouping of the last column cells in heatmaps in ps files. I used to prefer the ps because they embed more easily in latex documents (although pdf are not difficult and conversions are trivial anyhow), but I'm curious if there are other reasons why one format might be preferred over the other in this context. This is with R 2.6 on linux, and I've seen this behavior with older R version also. Francois sessionInfo() R version 2.6.0 (2007-10-03) x86_64-unknown-linux-gnu locale: LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C attached base packages: [1] stats graphics grDevices utils datasets methods base loaded via a namespace (and not attached): [1] rcompgen_0.1-15
ps or pdf
3 messages · Francois Pepin, Brian Ripley
On Mon, 31 Mar 2008, Francois Pepin wrote:
Hi everyone, I have been making a fair amount of figures in R recently that I've been touching up with Illustrator and I've found a difference between pdf and ps files and I was wondering if someone could enlighten me about them. While the figures look the same, the ps version tends to have truncated strings. The last character of short strings tends to be on a string of its own, located right beside the rest. This makes it a bit awkward to manipulate, especially if scaling is involved. Is there a reason for this differences?
Please see the footer of this message. Neither postscript() nor pdf() graphics devices split up strings they are passed (by e.g. text()), so this is being done either by the code used to create the plot (and we have no idea what that is) or by the viewer. I suspect the problem is rather in the viewer, but without the example we asked for it is impossible to know.
There also seems to be somewhat arbitrary grouping of the last column cells in heatmaps in ps files.
Again, we need an example.
I used to prefer the ps because they embed more easily in latex documents (although pdf are not difficult and conversions are trivial anyhow), but I'm curious if there are other reasons why one format might be preferred over the other in this context.
The graphics devices are very similar (they share a lot of code). One small difference is that PostScript has an arc primitive, and PDF does not.
This is with R 2.6 on linux, and I've seen this behavior with older R version also.
Nothing has changed at that level for a long time -- not even in current versions of R (and 2.6.0 is obsolete).
Francois sessionInfo() R version 2.6.0 (2007-10-03) x86_64-unknown-linux-gnu locale: LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C attached base packages: [1] stats graphics grDevices utils datasets methods base loaded via a namespace (and not attached): [1] rcompgen_0.1-15
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
Prof Brian Ripley wrote:
Please see the footer of this message.
Sorry, here is an example. For some reason, I cannot reproduce it
without using actual gene names.
set.seed(1)
##The row names were originally obtained using the hgug4112a library
##from bioconductor. I set it manually for people who don't have it
##installed.
##library(hgug4112a);row<-sample(na.omit(unlist(as.list(hgug4112aSYMBOL))),50)
row<-c("BDNF", "EMX2", "ZNF207", "HELLS", "PWP1", "PDXDC1", "BTD",
"NETO1", "SLCO4C1", "FZD7", "NICN1", "TMSB4Y", "PSMB7", "CADM2",
"SIRT3", "ADH6", "TM6SF1", "AARS", "TMEM88", "CP110", "ADORA2A",
"ATAD3A", "VAPA", "NXPH3", "IL27RA", "NEBL", "FANCF", "PTPRG",
"HSU79275", "CCDC34", "EPDR1", "FBLN1", "PCAF", "AP1B1", "TXNRD2",
"MUC20", "MBNL1", "STAU2", "STK32C", "PPIAL4", "TGFBR2", "DPY19L2P3",
"TMEM50B", "ENY2", "MAN2A2", "ZFYVE26", "TECTA", "CD55", "LOC400794",
"SLC19A3")
postscript('/tmp/heatmap.ps',paper='letter',horizontal=F)
heatmap(matrix(rnorm(2500),50),labRow=row)
dev.off()
Neither postscript() nor pdf() graphics devices split up strings they are passed (by e.g. text()), so this is being done either by the code used to create the plot (and we have no idea what that is) or by the viewer. I suspect the problem is rather in the viewer, but without the example we asked for it is impossible to know.
Example of row names that are truncated in Illustrator (* denoting truncation): CCDC3*4 (2nd row) MUC2*0 (3rd row) MBNL*1 (8th row) ... It is likely that Illustrator (CS 3, OS X version) is at fault. I do not see any truncation if I look at the ps file by hand (lines 4801 and 4802): 540.22 545.88 (MUC20) 0 0 0 t 540.22 553.90 (CCDC34) 0 0 0 t
There also seems to be somewhat arbitrary grouping of the last column cells in heatmaps in ps files.
Again, we need an example.
The top right cell (26, TXNRD2) is grouped with the cell just below it (26, CCDC34). It's more of a curiosity than anything else.
I used to prefer the ps because they embed more easily in latex documents (although pdf are not difficult and conversions are trivial anyhow), but I'm curious if there are other reasons why one format might be preferred over the other in this context.
The graphics devices are very similar (they share a lot of code). One small difference is that PostScript has an arc primitive, and PDF does not.
This is what I thought at first, which is why I found these differences surprising. I think your idea of blaming the viewer is correct. I thought that Adobe of all people could deal with Postscript files properly, but I guess I was overly trusting. Thanks for the help, Francois