Re: Bug in row_number() optimization

On Mon, 28 Nov 2022 at 22:59, Sergey Shinderuk
<s(dot)shinderuk(at)postgrespro(dot)ru> wrote:
>
> On 28.11.2022 03:23, David Rowley wrote:
> > On Sat, 26 Nov 2022 at 05:19, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >> It's pretty unlikely that this would work during an actual index scan.
> >> I'm fairly sure that btree (and other index AMs) have hard-wired
> >> assumptions that index operators are strict.
> >
> > If we're worried about that then we could just restrict this
> > optimization to only work with strict quals.
>
> Not sure this is necessary if btree operators must be strict anyway.

I'd rather see the func_strict() test in there. You've already
demonstrated you can get wrong results with a non-strict operator. I'm
not disputing that it sounds like a broken operator class or not. I
just want to ensure we don't leave any holes open for this
optimisation to return incorrect results.

David