Re: should we have a fast-path planning for OLTP starjoins?

On Mon, Feb 10, 2025 at 5:35 PM Tomas Vondra <tomas@vondra.me> wrote:
> On 2/10/25 08:29, Richard Guo wrote:
> > Hmm, I'm still a little concerned about whether the resulting joins
> > are legal.  Suppose we have a join pattern like the one below.
> >
> >  F left join
> >   (D1 inner join T on true) on F.b = D1.b
> >   left join D2 on F.c = D2.c;
> >
> > For this query, the original joinlist is [F, D1, T, D2].  If we
> > reorder it to [[F, T], D1, D2], the sub-joinlist [F, T] would fail to
> > produce any joins, as the F/T join is not legal.
> >
> > This may not be the pattern we are targeting.  But if we intend to
> > support it, I think we need a way to ensure that the resulting joins
> > are legal.

> It's quite possible the PoC patch I posted fails to ensure this, but I
> think the assumption is we'd not reorder joins for dimensions that any
> any join order restrictions (except for the FK join).

Then, we'll need a way to determine if a given relation has join-order
restrictions, which doesn't seem like a trivial task.  We do have the
has_join_restriction() function, but it considers any relations
involved in an outer join as having join restrictions, and that makes
it unsuitable for our needs here.

Thanks
Richard