Martijn van Oosterhout <kleptog(at)svana(dot)org> writes:
> This would mean that we wouldn't be assuming that tuples near the end
> take as long as tuples near the beginning. Except we're now dealing
> will smaller numbers, so I'm worried about error accumlation.
Hm, that would explain why Hash joins suffer from this especially. Even when
functioning properly hashes get slower as the buckets fill up and there are
longer lists to traverse. Perhaps the hashes are suffering inordinately from
collisions though. Some of the data type hash functions looked kind of suspect
when I peeked at them a while back.
--
greg