note sk Very interesting! Thanks for the new idea!<p> I lost my server connection for some reason and so I tested this on my laptop and I do see a very good improvement with the your modification.<p> <b> update: corrected <> with $_ per pijll post</b> <code> C:\>perl -lne "BEGIN{$,=','} print+(split',',$_)[0..14] " > junk </code> this finishes in about 14 seconds....<b> corrected timing</b><p> <code> C:\>perl -lanF, -e "BEGIN{ $,=\",\"} print @F[0..14];" numbers.csv > junk </code> this takes about 18 seconds<p> I don't have a timing utility in Windows so the times are just wallclock times.<p> I guess windows is faster because the process run at 100% CPU (or whatever is required i guess?). On the UNIX servers the process might be more time-shared?<p> My laptop is 1.6G Centrino/1GB Ram/perl, v5.6.1<p> cheers<p> SK <b>Update:</b> Thanks pijll, the time it takes to run your version of the code is almost same as the one that uses -n. 448559 448598