Another question: are the second and third fields of each record important or can they be thrown away?

Assuming they are unimportant, I suggest you use a hash of arrays (if you want to keep all fourth fields, even if they are repeated) or a hash of hashes (if you want to collapse repeated fourth fields into one item).

CountZero

"If you have four groups working on a compiler, you'll get a 4-pass compiler." - Conway's Law