Rob,
> q5 and q6 are too complex to discuss here, but the fundamental issue
> is the order in which postgres decides to do things. The choice for
> me is clear: the developer time trying to figure out how to make the
> planner do the "obviously right thing" has been too high with
> postgres. These tests demonstate to me that for even complex queries,
> oracle wins for our problem.
>
> It looks like we'll be migrating to oracle for this project from these
> preliminary results. It's not just the planner problems. The
> customer is more familiar with oracle, and the vacuum performance is
> another problem.
Hey, we can't win 'em all. If we could, Larry would be circulating his
resume'.
I hope that you'll stay current with PostgreSQL developments so that you can
do a similarly thourough evaluation for your next project.
--
-Josh Berkus
Aglio Database Solutions
San Francisco