in reply to Re^2: parsing a very large array with regexps
in thread parsing a very large array with regexps

Oh. I think I see the problem. How many patterns are you searching for, what do the patterns actually look like, and how long is "too long"? With a 2 GHz Pentium IV, I can search a 2.2 Mb file with 274K lines for three five character patterns in about 1.35 seconds, and thirteen in 2.00 seconds (each additional pattern adds around .065 seconds or so.)

Those numbers are for simple patterns like /abcde|bdfhj|acegi/. If your patterns are more like

/a[bc]+d|[^efg](h|i)|klm*n+/
each pattern will might take on the order of ten times as long.

If you have any pathological patterns with nested * or + qualifiers, things can get very slow indeed.