[valenterry]

Postgres natively supports UUID, so I'm using that and guess that it will internally optimize things. Otherwise I would certainly use a different format.