Re: Virtual generated columns

On Fri, 14 Feb 2025 at 10:59, Peter Eisentraut <peter(at)eisentraut(dot)org> wrote:
>
> On 13.02.25 14:06, jian he wrote:
> > I didn't solve the out join semantic issue.
> > i am wondering, can we do the virtual generated column expansion in
> > the rewrite stage as is,
> > and wrap the expressions in PHVs if the virtual generated
> > columns come from the nullable side of an outer join.
>
> PlaceHolderVar looks like a fitting mechanism for this. But it's so far
> a planner node, so it might take some additional consideration if we
> want to expand where it's used.

It seems pretty baked into how PHVs work that they should only be
added by the planner, so I think that I agree with Richard -- virtual
generated columns probably have to be expanded in the planner rather
than the rewriter.

> Maybe a short-term fix would be to error out if we find ourselves about
> to expand a Var with varnullingrels != NULL. That would mean you
> couldn't use a virtual generated column on the nullable output side of
> an outer join, which is annoying but not fatal, and we could fix it
> incrementally later.

I think that would be rather a sad limitation to have. It would be
nice to have this fully working for the next release.

Attached is a rough patch that moves the expansion of virtual
generated columns to the planner. It needs a lot more testing (and
some regression tests), but it does seem to fix all the issues
mentioned in this thread.

Regards,
Dean