Re: Discussion on missing optimizations

On 2017-10-08 17:11:44 -0400, Tom Lane wrote:
> Andres Freund <andres(at)anarazel(dot)de> writes:
> > On 2017-10-08 11:28:09 -0400, Tom Lane wrote:
> >> https://commitfest.postgresql.org/15/1001/
> >> The reason that's not in v10 is we haven't been able to convince
> >> ourselves whether it's 100% correct.
>
> > Unfortunately it won't help in this specific case (no support for UNION,
> > just UNION ALL), but I thought it might be interesting to reference
> > https://medium.com/@uwdb/introducing-cosette-527898504bd6
> > here.
>
> Huh, that is an interesting project indeed. Although I'm not sure that
> it quite addresses the question of whether an optimization transform
> is valid. IIUC, it could prove that a particular query having been fed
> through the transform didn't change semantics, but that offers only
> limited insight into whether some other query fed through the transform
> might change.

According to the guide it offers some support for more general
transformations:
http://cosette.cs.washington.edu/guide#24-symbolic-predicates That's
still only going to be sufficient for some of the interesting cases, but
still...

Wonder about pinging them about the OR -> UNION case, they've been
responsive to problem in some threads I found online.

Greetings,

Andres Freund