It seems like the hash lookup should be faster than doing a bunch of regex matches, but can't you just throw some examples at it and time the two?