Re: Performing partition pruning using row value

On Fri, Jul 10, 2020 at 9:35 AM Etsuro Fujita <etsuro(dot)fujita(at)gmail(dot)com> wrote:
> On Thu, Jul 9, 2020 at 7:57 PM Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com> wrote:
> > On 2020/07/09 19:45, Etsuro Fujita wrote:
> > > Please add the patch to the next CF so that it does not get lost.
> >
> > Is this a bug rather than new feature?
>
> I think it's a limitation rather than a bug that partition pruning
> doesn't support row-wise comparison, so I think the patch is a new
> feature.

I tend to think so too. IMO, partition pruning, like any other
optimization, works on a best-effort basis. If the result it produces
is wrong, now that would be a bug, but I don't think that's the case
here. However, I do think it was a bit unfortunate that we failed to
consider RowCompare expressions when developing partition pruning
given, that index scans are already able to match them.

Speaking of which, I hope that Kato-san has looked at functions
match_rowcompare_to_indexcol(), expand_indexqual_rowcompare(), etc. in
indxpath.c as starting points for the code to match RowCompares to
partition keys.

--
Amit Langote
EnterpriseDB: http://www.enterprisedb.com