in reply to Moving from hashing to tie-ing.

I've found that using Tie::File for large files (like yours) has not given me the performance I was seeking. Is it possible for you to divide your process into two steps, one to parse the file and store the relevant components into a relational database, and the second to process the various components according to your needs, using the power of the relational database? I've had good success with Oracle and MySQL ...

Update: Yeah, like Fletch said. Curse those johnny-on-the-spot monks! :)


No good deed goes unpunished. -- (attributed to) Oscar Wilde