Re: GEQO vs join order restrictions

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> Tom, do you think the "independent subproblem" stuff from last night
> would be worth pursuing?

It's worth looking into. I'm not certain if it will end up being a good
idea or not. Right now the joinlist collapse code is pretty stupid
(as you know --- it only thinks about the collapse_limit variables,
plus IIRC it knows about FULL JOIN). Making it smarter might result in
duplication of logic, or require unpleasant refactoring to avoid such
duplication, or even add more cycles than it's likely to save later on.
Another issue is order of operations: I'm not sure if all the
information needed to make such decisions has been computed at that
point. But we won't know unless we try it. It seems at least
potentially useful.

regards, tom lane