Dear R helpers,
   For me ( i.e. R 2.1.1 on Mac OS X), using  "trellis.device 
(postscript, onefile = F, etc ..."  with the lattice library within a R 
function works fine to obtain the desired graph as an EPS file , 
provided that :
    1) the command dev.off() is not included in this function
    2) and it is  issued at the  command level after the function has 
been exited
I would like to know if there is a way to close the EPS file within the 
function itself, freeing the user to issue the closing command (I 
already  tried trellis.device (), and trellis.device (null) without any 
success).
Regards,
J.-M.
  ----
Jean-Marc Ottorini               LERFoB, UMR INRA-ENGREF 1092
  email  ottorini at nancy.inra.fr          INRA - Centre de Nancy
  voice  +33-0383-394046                    F54280 - Champenoux
  fax    +33-0383-394034                                 France
Prof Brian Ripley
2005-Aug-26  14:36 UTC
[R] Shutting down a trellis plot (was R-help Digest, Vol 30, Issue 26)
I suspect you have not print()-ed your graphics, see FAQ Q7.22.
It is then possible to include dev.off() within the function.  E.g.
testit <- function(fn = "test.eps")
{
   trellis.device(postscript, file=fn, onefile = FALSE, horizontal=FALSE)
   print(stripplot(voice.part ~ jitter(height), data = singer, aspect = 1,
                   jitter = TRUE, xlab = "Height (inches)"))
   dev.off()
}
testit()
works for me.
On Fri, 26 Aug 2005, Jean-Marc Ottorini wrote:
>   For me ( i.e. R 2.1.1 on Mac OS X), using  "trellis.device
> (postscript, onefile = F, etc ..."  with the lattice library within a
R
> function works fine to obtain the desired graph as an EPS file ,
> provided that :
>
>    1) the command dev.off() is not included in this function
>
>    2) and it is  issued at the  command level after the function has
> been exited
>
> I would like to know if there is a way to close the EPS file within the
> function itself, freeing the user to issue the closing command (I
> already  tried trellis.device (), and trellis.device (null) without any
> success).
-- 
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