Just one comment on sort efficiency: contrary to what a couple of posters seem to think, Perl sorts strings faster than numbers. The whole point of Uri and Larry's paper is that as string sort is the default, hard-coded sort, it does not have to call any Perl subroutine, and thus works _real_ fast. So the trick to fast sorting is to be able to do just a sort @array, not a sort {$a <=> $b}, @array.
And BTW if you are looking for a module to do this automagically, well it looks like File::Records is looking for a new owner.