Think of your csv-file as a simple form of database and then you can use DBI and DBD::CSV on it and use the strength of SQL to select and sort as you like.
CountZero
"If you have four groups working on a compiler, you'll get a 4-pass compiler." - Conway's Law