[R] Efficiently reading random lines form a large file

Marc Schwartz marc_schwartz at comcast.net
Wed May 16 01:19:08 CEST 2007


On Tue, 2007-05-15 at 16:02 -0700, Juan Pablo Lewinger wrote:
> I need to read two different random lines at a time from a large 
> ASCII file (120 x 296976) containing space delimited 0-1 entries.
> 
> The following code does the job and it's reasonable fast for my needs:
> 
>    lineNumber = sample(120, 2)
>    line1 = scan(filename, what = "integer", skip=lineNumber[1]-1, nlines=1)
>    line2 = scan(filename, what = "integer", skip=lineNumber[2]-1, nlines=1)
> 
>  > system.time(for (i in 50){
> +   lineNumber = sample(120, 2)
> +   line1 = scan(filename, what = "integer", skip=lineNumber[1]-1, nlines=1)
> +   line2 = scan(filename, what = "integer", skip=lineNumber[2]-1, nlines=1)
> + })
> 
> Read 296976 items
> Read 296976 items
> [1] 14.24  0.12 14.51    NA    NA
> 
> However, I'm wondering if there's an even faster way to do this. Is there?

You might want to take a look at this post by Jim Holtman from earlier
in the year for some ideas:

http://tolstoy.newcastle.edu.au/R/e2/help/07/02/9709.html

HTH,

Marc Schwartz



More information about the R-help mailing list