Well, is Perl fast enough for the job? If not (your call, I guess), you'll have to code the equivalent functionality in C, which should give you a bit of an improvement in speed.

What machine are you running this on? How long does the script take to run? I can't remember any benchmarks, but I know I've processed large files before and Perl whizzes through them. What you've discussed is not any heavy processing, so I can't imagine that your script would take more than a few minutes to process 1,000,000 lines.