Joe Conway wrote:
> regression=# SELECT * FROM connectby('connectby_int', 'keyid',
> 'parent_keyid', '2', 0, '~') AS t(keyid int, parent_keyid int, level
> int, branch text);
Yeah, I saw that. But I don't think I can use it for a few reasons:
- I could't see a way to do multi-item sorts, like by score descending
and then post time ascending, for example.
- I couldn't see a way to filter for "messageroot_id = ?" first, so it
would probably be unnecesarily slow as the number of messages in this
discussion / the number of messages total ratio becomes small.
- Also performance - I can't cache the results as a row in the table,
which I could see being important at some point. Kind of a pain to keep
in sync, like all denormalized things, but I could see this being an
expensive operation.
Thanks,
Scott
