R help - May 2005 - efficient filtering of matrices

I was wondering if someone can tell me the best way to search through a 
matrix and choose certain rows (based on certain conditions) to put into a 
separate matrix.
What I have tried so far is very slow for a large dataset I'm working with. 
e.g., I have this piece of code to create a new matrix (newmat) based on my 
filtering conditions.  Do I need to do this kind of thing where I keep 
rbinding?

newmat<-rep(NA,12)
for (i in 1:length(origmat[,1])
{
	if (  is.na(origmat[i,10])  |  (!is.na(origmat[i,10]) & (origmat[i,2] <=
origmat[i,10]) )   )
		newmat<-rbind(newmat, origmat[i,])
}


Thanks,
Steven

Yikes, that looks complicated. I think what you're trying to do is available
already, see:

?subset

-- 
Tyler Smith

PhD Candidate
Plant Science Department
Faculty of Agricultural and Environmental Sciences
McGill University
MacDonald Campus
21,111 Lakeshore
Ste. Anne de Bellevue, Quebec, Canada H9X 3V9

Tel.: 514-398-7851 ext. 8726
Fax: 514-398-7897
tsmith20 at po-box.mcgill.ca

sms13+ at pitt.edu wrote:
> I was wondering if someone can tell me the best way to search through a 
> matrix and choose certain rows (based on certain conditions) to put into 
> a separate matrix.
> What I have tried so far is very slow for a large dataset I'm working 
> with. e.g., I have this piece of code to create a new matrix (newmat) 
> based on my filtering conditions.  Do I need to do this kind of thing 
> where I keep rbinding?
> 
> newmat<-rep(NA,12)
> for (i in 1:length(origmat[,1])
> {
>     if (  is.na(origmat[i,10])  |  (!is.na(origmat[i,10]) & 
> (origmat[i,2] <= origmat[i,10]) )   )
>         newmat<-rbind(newmat, origmat[i,])
> }

What about vectorizing the stuff?

   index <- is.na(origmat[ ,10]) |
     (!is.na(origmat[ ,10]) &
       (origmat[ ,2] <= origmat[ ,10]))
   newmat <- origmat[index,]

Uwe Ligges

> 
> Thanks,
> Steven
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html