[R] importing large objects

Sirotkin, Alexander demiurg at ti.com
Sat Jun 15 23:29:03 CEST 2002


I see you are extremely angry at TI for some reason. 

Well, first and foremost : I do not use R at TI, I use it for 
my master thesis project at Tel-Aviv University. I also happen 
to be a TI employee, but I don't think it matters. Maybe I should 
use my University mail account while posting such questions, but
I just never encountered such kind of hostility before.

As far as I know, nobody at TI uses R, so I don't think I can 
convince the TI management to donate anything to R project.

I do have enough memory (I hope 512M is enough), the question
actually was : is there a limit (hard or soft, by soft meaning that 
the performance degrades significantly after object size reachs some
limit)
on objct size in R. If there is not (meaning it can handle large objects
and 
it scales well when object sizes increase) - then I'm apparently doing
something wrong and I should revist my code. If not - I should switch to
S-plus.

P.S. you can reply to my university account ( demiurg at tau.ac.il
<mailto:demiurg at tau.ac.il> ) if it makes 
you happier.

Thanks for your help.

ripley at stats.ox.ac.uk <mailto:ripley at stats.ox.ac.uk>  wrote:


On Sat, 15 Jun 2002, Sirotkin, Alexander wrote:



  

I was trying to import a very large data table to R using RODBC. By

large meaning a table of 600 real

numbers and 10000 rows. It failed saying that it reached a memory limit.

    



Which is what?  Buy some more memory: I am sure TI can afford it!



  

This is extremely weird

because S-Plus can import the same table, and it does this quite fast.

    



Why is it `extremely wierd'?  Why do you expect user-contributed
software

(which is not even in the main CRAN section) for a free package to be as

memory-efficient as a commercial package?



  

Are there any limitations on objects size in current implementation of R
?

    



Yes, around 2Gb for each object (increased in the current R-devel).



We would welcome contributions of improvements to RODBC or funds to

develop it (and I note that you are posting from a large commercial

concern).



  



-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://stat.ethz.ch/pipermail/r-help/attachments/20020615/c687e29f/attachment.html


More information about the R-help mailing list