Re: Parallel INSERT (INTO ... SELECT ...)

On Mon, Jan 18, 2021 at 3:50 PM Greg Nancarrow <gregn4422(at)gmail(dot)com> wrote:
>
> On Mon, Jan 18, 2021 at 8:10 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
> >
> > > >It is not clear why the above two are shouldn't happen cases and if so
> > > >why we want to treat them as unsafe. Ideally, there should be an
> > > >Assert if these can't happen but it is difficult to decide without
> > > >knowing why you have considered them unsafe?
> > >
> > > The checks being done here for "should never happen" cases are THE
> > > SAME as other parts of the Postgres code.
> > > For example, search Postgres code for "null conbin" and you'll find 6
> > > other places in the Postgres code which actually ERROR out if conbin
> > > (binary representation of the constraint) in a pg_constraint tuple is
> > > found to be null.
> > > The cases that you point out in the patch used to also error out in
> > > the same way, but Vignesh suggested changing them to just return
> > > parallel-unsafe instead of erroring-out, which I agree with.
> > >
> >
> > You have not raised a WARNING for the second case.
>
> The same checks in current Postgres code also don't raise a WARNING
> for that case, so I'm just being consistent with existing Postgres
> code (which itself isn't consistent for those two cases).
>

Search for the string "too few entries in indexprs list" and you will
find a lot of places in code raising ERROR for the same condition.

--
With Regards,
Amit Kapila.