Comment by jonesnc

Why does the randomization have to happen in the database query? Assuming there aren't any large gaps in the distribution of IDs, and if you know the max_id, couldn't you pick a random number between min_id and max_id and get the record whose ID matches that random number?