you're right, although I don't think compression would save a ton of space since the hashes are random. Might save a bit or maybe 2 per entry