[R] how to control significant digits(?) on axis labels

[Paul.Rustomji at csiro.au]

Dear R mailing list

I would like some help on how to get R to display the same number of significant digits (?) for *all* tick marks on axis labels (yet be flexible enough to handle different data sets that vary by 10-1000X).

Consider this simple example:

#---------------------------#
x <- seq(0,2,by=0.5)
xlim <- range(pretty(x,n=5))
ylim <- xlim

plot(c(xlim[1],xlim[2])~c(ylim[1],ylim[2]),type="n",xlim=xlim,ylim=ylim,xaxt="n",yaxt="n",lty=1,pch=40,xaxs="i",yaxs="i",xlab="",ylab="")

par(xaxt="s")
par(xaxs="i")
at <- pretty(xlim)
print(at)
axis(1,at=at,labels=at,las=1,tcl=-0.4,lwd=0.8,cex.axis=1,padj=-1)
mtext("x",1,line=1.8,cex=1,padj=-0)

par(yaxt="s")
par(yaxs="i")
at <- pretty(ylim)
print(at)
axis(2,at=at,labels=at,las=1,tcl=-0.4,lwd=0.8,cex.axis=1,padj=0.3,hadj=0.6)
mtext("y",2,line=2.1,cex=1,padj=-0.5,las=3)

points(c(xlim[1],xlim[2])~c(ylim[1],ylim[2]),type="l",lwd=0.4)
points(x,x,col="red")

#---------------------------#

This results in axis labels of 

0, 0.5, 1, 1.5, 2 when I would like 

0.0, 0.5, 1.0, 1.5, 2.0.


The wierd thing is the print(at) statements above show what I want (ie a consistent number of decimal places) but somewhere in the axis labelling procedure things seem to change.

I have tried using
> sprintf("%.1f",at)

which works for the range of data above, but this then gives unsatisfactory results if the data is 1+ order of magnitude greater (which it could be in the data I need this to work on) in which case the decimal place is best avoided (eg I would like labels of  0, 5, 10, 15, 20 etc rather than 0.0, 5.0, 15.0, 20.0, 25.0)

Also I think signif() wouldn't be quite what I want and have tried mucking around with formatC but to no avail.

Any ideas?
Paul.


> sessionInfo()
R version 2.8.1 (2008-12-22) 
i386-pc-mingw32 

locale:
LC_COLLATE=English_Australia.1252;LC_CTYPE=English_Australia.1252;LC_MONETARY=English_Australia.1252;LC_NUMERIC=C;LC_TIME=English_Australia.1252

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] xlsReadWrite_1.3.3 qualV_0.2-4        KernSmooth_2.22-22

loaded via a namespace (and not attached):
[1] tools_2.8.1


Paul Rustomji
Rivers and Estuaries
CSIRO Land and Water
GPO Box 1666
Canberra ACT 2601

ph +61 2 6246 5810
mobile 0406 375 739