Horton

Thread

Commits

  1. Don't pass down nonnullable_vars while reducing outer joins.

  1. Do we need to pass down nonnullable_vars when reducing outer joins?

    Richard Guo <guofenglinux@gmail.com> — 2022-09-13T07:00:16Z

    AFAICS, the Vars forced nonnullable by given clause are only used to
check if we can reduce JOIN_LEFT to JOIN_ANTI, and it is checking the
join's own quals there. It seems to me we do not need to pass down
nonnullable_vars by upper quals to the children of a join.

Attached is a patch to remove the pass-down of nonnullable_vars.

Thanks
Richard

  2. Re: Do we need to pass down nonnullable_vars when reducing outer joins?

    Tom Lane <tgl@sss.pgh.pa.us> — 2022-11-05T20:00:47Z

    Richard Guo <guofenglinux@gmail.com> writes:
> AFAICS, the Vars forced nonnullable by given clause are only used to
> check if we can reduce JOIN_LEFT to JOIN_ANTI, and it is checking the
> join's own quals there. It seems to me we do not need to pass down
> nonnullable_vars by upper quals to the children of a join.

Hmm, you are right, we are not doing anything useful with that data.
I can't remember if I had a concrete plan for doing something with it
or not, but we sure aren't using it now.  So pushed.

			regards, tom lane

  3. Re: Do we need to pass down nonnullable_vars when reducing outer joins?

    Richard Guo <guofenglinux@gmail.com> — 2022-11-07T02:54:32Z

    On Sun, Nov 6, 2022 at 4:00 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:

> Richard Guo <guofenglinux@gmail.com> writes:
> > AFAICS, the Vars forced nonnullable by given clause are only used to
> > check if we can reduce JOIN_LEFT to JOIN_ANTI, and it is checking the
> > join's own quals there. It seems to me we do not need to pass down
> > nonnullable_vars by upper quals to the children of a join.
>
> Hmm, you are right, we are not doing anything useful with that data.
> I can't remember if I had a concrete plan for doing something with it
> or not, but we sure aren't using it now.  So pushed.


Thanks for pushing it!

Thanks
Richard