Re: Join removal and attr_needed cleanup

2025-05-20 Thread Bennie Swart
18:17, Tom Lane wrote: Bennie Swart writes: -- join elimination fails -- expect both b and c to be eliminated, but b remains explain (costs off)   select a.*     from foo a   left join foo b on (b.id1, b.id2) = (a.id1, a.id2)   left join foo c on (c.id1, c.id2)

Re: Join removal and attr_needed cleanup

2024-11-10 Thread Bennie Swart
We are encountering this issue which results in poor planning for some views. Some examples to illustrate the issue: -- setup create table foo as   select id1, id2     from generate_series(1, 100) id1, generate_series(1, 100) id2; alter table foo add unique (id1, id2); -- join elimina

Re: Wrong security context for deferred triggers?

2024-07-01 Thread Bennie Swart
straint trigger is deferred, but works as expected when it is immediate. Hopefully this helps to better paint the picture. Our workaround currently is to just make `legacy.stock_check_state()` security definer as well. As I told Laurenz, we're not looking to advocate for any specific outcom