As long as the files aren't large, that's fine, except that you're not finding keys present in file B but not in file A.

You're using 2 hashes where one would suffice, so you can cut your memory usage in half.

Pseudo code:

Process File A contents and store in %hash_one

while(<FileB>) {
    # do whatever you need to to compute $key
    if (! exists $hash_one{$key}) {
        print "File B contains $key but File A does not\n";
    }
    delete $hash_one{$key};
}

# the keys left in %hash_one weren't in File B
foreach(keys %hash_one) {
    print "File A contains $key, but File B does not\n";
}
[download]