Re: Redundant Result node - Mailing list pgsql-hackers

From David Rowley
Subject Re: Redundant Result node
Date
Msg-id CAApHDvo14gwvtSdhYzn=6CdUSaVpnMPwUct4Qx4j+fO+5TixoA@mail.gmail.com
Whole thread Raw
In response to Redundant Result node  (Richard Guo <guofenglinux@gmail.com>)
List pgsql-hackers
On Thu, 22 Aug 2024 at 19:34, Richard Guo <guofenglinux@gmail.com> wrote:
>     /* Add projection step if needed */
>     if (sorted_path->pathtarget != target)
>         sorted_path = apply_projection_to_path(root, ordered_rel,
>                                                sorted_path, target);
>
> This does not seem right to me, as PathTargets are not canonical, so
> we cannot guarantee that two identical PathTargets will have the same
> pointer.  Actually, for the query above, the two PathTargets are
> identical but have different pointers.
>
> I wonder if we need to invent a function to compare two PathTargets.
> Alternatively, in this case, would it suffice to simply compare
> PathTarget.exprs?

I think tlist.c would be a good home for such a function. If you go
with the function route, then it's easier to add optimisations such as
checking if the pointers are equal before going to the trouble of
checking if the exprs match.

David



pgsql-hackers by date:

Previous
From: Jakub Wartak
Date:
Subject: Re: Enable data checksums by default
Next
From: Melih Mutlu
Date:
Subject: Re: ANALYZE ONLY