[R] dev.off() inside a function & other glitches

[Michael Hopkins]

Thanks to all who replied on list and to me directly.

You are right, FAQ 7.22 is the answer.  For any others who get caught  
by this it's important to print(..) a plot with a function, not just  
create it, i.e.

	print( xyplot(...   ) )

not just

	xyplot(... )

..otherwise it doesn't necessarily get sent to the opened device.

Michael


On 2 Jul 2008, at 01:06, Deepayan Sarkar wrote:

> On 7/1/08, Michael Hopkins <michael.hopkins at hopkins-research.com>  
> wrote:
>>
>>
>> Hi R people
>>
>> I am using a function to create a pdf device, then send a lot of  
>> plots
>> to it in a loop then a last lattice xyplot (itself within a function)
>> outside the loop and finally call dev.off() to write to the file.
>> This works well apart from the fact that the last plot does not get
>> sent to the file unless I comment out dev.off() and then apply it in
>> the console afterwards instead:
>>
>>        plot_stuff( ...); dev.off()
>>
>> The device is opened like this:
>>
>>        pdf( paste( var_string, ".pdf", sep="" ), onefile=TRUE,  
>> paper="a4r",
>> width=9, height=6.5 )
>>
>> Also, if I try to send two different xyplots after the loop only the
>> last one ever gets written to the file, whether or not I apply the
>> dev.off() trick above.
>>
>> Any thoughts on why this stuff happens and best ways to avoid it are
>> appreciated.
>
> This sounds like FAQ 7.22:
>
> http://cran.us.r-project.org/doc/FAQ/R-FAQ.html#Why-do-lattice_002ftrellis-graphics-not-work_003f
>
> The dev.off() issue is most likely a red herring; the important bit is
> that the xyplot() call was not the last expression in your function.
>
> -Deepayan
>



                             Web:    http://www.hopkins-research.com/
                          Office:    +44 (O)1732 228844
                          Mobile:    +44 (O)7813 467381