[R] a function more appropriate than 'sapply'?

[Uwe Ligges]

On 26.01.2013 20:46, Berend Hasselman wrote:
>
> On 26-01-2013, at 19:43, emorway <emorway at usgs.gov> wrote:
>
>> I'm wondering if I need to use a function other than sapply as the following
>> line of code runs indefinitely (or > 30 min so far) and uses up all 16Gb of
>> memory on my machine for what seems like a very small dataset (data attached
>> in a txt file  wells.txt
>> <http://r.789695.n4.nabble.com/file/n4656723/wells.txt>  ).  The R code is:
>>
>> wells<-read.table("c:/temp/wells.txt",col.names=c("name","plc_hldr"))
>> wells2<-wells[sapply(wells[,1],function(x)length(strsplit(as.character(x),
>> "_")[[1]])==2),]
>>
>> The 2nd line of R code above gets bogged down and takes all my RAM with it:
>> <http://r.789695.n4.nabble.com/file/n4656723/memory_loss.png>
>>
>> I'm simply trying to extract all of the lines of data that have a single "_"
>> in the first column and place them into a dataset called "wells2".  If that
>> were to work, I then want to extract the lines of data that have two "_" and
>> put them into a separate dataset, say "wells3".  Is there a better way to do
>> this than the one-liner above?
>
>
> Read your file with
>
> 	wells<-read.table("wells.txt",col.names=c("name","plc_hldr"), stringsAsFactors=FALSE)
>
> Remove all non underscores with
>
> 	w.sub <- gsub("[^_]+","",wells[,1])
>
> then select elements of w.sub with 2 underscores and a single underscore with
>
> 	u.2 <- which(w.sub=="__")
> 	u.1 <- which(w.sub=="_")
>
> and use u.1 and u.2 to select the appropriate rows of wells.

With grep:

wells1 <- wells[grep("^[^\\_]*_[^\\_]*$", wells[,1]),]
wells2 <- wells[grep("^[^\\_]*_[^\\_]*_[^\\_]*$", wells[,1]),]


Best,
Uwe Ligges


> I tried to select rows containing 1 or 2 underscores with grep regular expressions but that appeared to be more difficult than I had expected.
> The method above is quick.
>
> Berend
>
> ______________________________________________
> 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.
>