This is not rather the native perl silution, but i solved such problem by writing hashes to plain text files and comparing them using sort and (if needed) uniq system functions and then compared two files with cmp (assuming you are running *nix systems).
This works much faster by all means (I used it for Apache logs proccessing)
These are not native perl functions, but the value of perl, among others is its flexibility and ability to be integreated with other programs
--
With best regards
Maqs.