R help - Sep 2009 - Any concept of 'by reference' or 'address of' in R?

Hi,
   Is there a way to pass the address of a data.frame through a set of
functions in R? I've got some code which is slowing down I think
because my data.frames are getting much larger - now approaching 1
million rows by 50-100 columns - and my functions - originally written
for much smaller data elements are sampling these data.frames a couple
of layers down.

   Before I rewrite them to do the sampling at the top level if there
was a way to pass the address of the data.frame instead of the a copy
of the whole thing then I suspect that would speed things up
significantly.

Thanks,
Mark

You can use R environments or proto objects from the proto package
(http://r-proto.googlecode.com) which are environments with additional
oo support using the prototype model) or the oo.R package which is similar
but uses a more conventional oo model.

On Fri, Sep 18, 2009 at 7:38 PM, Mark Knecht <markknecht at gmail.com>
wrote:
> Hi,
> ? Is there a way to pass the address of a data.frame through a set of
> functions in R? I've got some code which is slowing down I think
> because my data.frames are getting much larger - now approaching 1
> million rows by 50-100 columns - and my functions - originally written
> for much smaller data elements are sampling these data.frames a couple
> of layers down.
>
> ? Before I rewrite them to do the sampling at the top level if there
> was a way to pass the address of the data.frame instead of the a copy
> of the whole thing then I suspect that would speed things up
> significantly.
>
> Thanks,
> Mark
>
> ______________________________________________
> 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.
>

On Sep 18, 2009, at 7:38 PM, Mark Knecht wrote:
> Hi,
>   Is there a way to pass the address of a data.frame through a set of
> functions in R? I've got some code which is slowing down I think
> because my data.frames are getting much larger - now approaching 1
> million rows by 50-100 columns - and my functions - originally written
> for much smaller data elements are sampling these data.frames a couple
> of layers down.
>
>   Before I rewrite them to do the sampling at the top level if there
> was a way to pass the address of the data.frame instead of the a copy
> of the whole thing then I suspect that would speed things up
> significantly.
>
You could pass <it>'s name as a string and then when it arrived at the
destination, ... get() it:

?get

-- 
David Winsemius, MD
Heritage Laboratories
West Hartford, CT

I wouldn't assume R is making a copy; it depends on what you're doing 
with the dataframe.

I'd suggest using Rprof to find out where the cpu time is being 
spent. If you can adapt your problem to use all numbers or all 
characters then you can store the data in a matrix instead of a data 
frame, and potentially speed things up. (That advice is based on my 
own experience from a fairly large number of years ago, so it's 
possible that developments in R since then could negate the advice.)

-Don

At 4:38 PM -0700 9/18/09, Mark Knecht wrote:
>Hi,
>    Is there a way to pass the address of a data.frame through a set of
>functions in R? I've got some code which is slowing down I think
>because my data.frames are getting much larger - now approaching 1
>million rows by 50-100 columns - and my functions - originally written
>for much smaller data elements are sampling these data.frames a couple
>of layers down.
>
>    Before I rewrite them to do the sampling at the top level if there
>was a way to pass the address of the data.frame instead of the a copy
>of the whole thing then I suspect that would speed things up
>significantly.
>
>Thanks,
>Mark
>
>______________________________________________
>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.

-- 
--------------------------------------
Don MacQueen
Environmental Protection Department
Lawrence Livermore National Laboratory
Livermore, CA, USA
925-423-1062

On Fri, Sep 18, 2009 at 8:11 PM, Don MacQueen <macq at llnl.gov>
wrote:
> I wouldn't assume R is making a copy; it depends on what you're
doing with
> the dataframe.
Indeed ... I thought R had copy-on-write semantics for function
arguments, no? If so, then if your function doesn't change the
elements in the matrix/data.frame, then you shouldn't have to worry,
and look to optimize elsewhere ...

-steve

-- 
Steve Lianoglou
Graduate Student: Computational Systems Biology
 | Memorial Sloan-Kettering Cancer Center
 | Weill Medical College of Cornell University
Contact Info: http://cbio.mskcc.org/~lianos/contact