> Based on past experience, from a bang-for-buck perspective, I'd probably do
> this in the numerical order. What do you think? I know what I like and can
> do but I don't really know enough about PostgreSQL's performance weaknesses
> yet.
>
> What are we getting killed on?
>
I'm not a developer, but one thing I see come up occasionally around here are
planner issues. Sometimes people get really hammered by the planner choices,
and aren't provided a very good way to tune it. If you were able to eliminate
some worst-case-scenario type situations, that would make the few people who
are having problems very happy (I remember one thread in particular seemed
nasty). If I remember correctly, some developers don't much like the idea of
query hints, and I don't blame them, so you might want to run your ideas by
them first.
Also, this kind of modification might require significant additions to the
statistics system. The planner might be smart, but if it doesn't have any
more information you might not be able to get any more out of it. Autovacuum
might help with that as well (i.e. the info will be more up to date).
Regards,Jeff Davis
