On Sat, Apr 26, 2025 at 11:04 PM Alexander Korotkov <aekorot...@gmail.com> wrote: > On Sun, Apr 6, 2025 at 12:02 AM Alena Rybakina > <a.rybak...@postgrespro.ru> wrote: > > Should we add more regression tests covering these cases? > > > > I experimented with some examples like this and noticed that it does affect > > cardinality estimation, though I'm not sure the impact is significant. > > I used the tables from the regression tests, so if they’re appropriate for > > reproducing this case, it should be straightforward to add them. > > Thank you for your feedback. I've check the cases you've provided. I > found that the differences here are related to the SJE itself, not to > changes regarding PHVs handling. I think it generally OK that > estimates are somewhat changed due to such significant query > transformation. Hopefully they should be improved in the majority of > cases. > > I did some improvements to PHVs patch: revised comments and commit > message. I'm going to push it if no objections.
Uh, v2 was there already. That should be v3. ------ Regards, Alexander Korotkov Supabase
v3-0001-Disallow-removing-placeholders-during-Self-Join-E.patch
Description: Binary data
