Re: [HACKERS] WIP: Aggregation push-down

Robert Haas <robertmhaas(at)gmail(dot)com> wrote:

> On Fri, Feb 23, 2018 at 11:08 AM, Antonin Houska <ah(at)cybertec(dot)at> wrote:
> > I spent some more time thinking about this. What about adding a new strategy
> > number for hash index operator classes, e.g. HTBinaryEqualStrategyNumber? For
> > most types both HTEqualStrategyNumber and HTBinaryEqualStrategyNumber strategy
> > would point to the same operator, but types like numeric would naturally have
> > them different.
> >
> > Thus the pushed-down partial aggregation can only use the
> > HTBinaryEqualStrategyNumber's operator to compare grouping expressions. In the
> > initial version (until we have useful statistics for the binary values) we can
> > avoid the aggregation push-down if the grouping expression output type has the
> > two strategies implemented using different functions because, as you noted
> > upthread, grouping based on binary equality can result in excessive number of
> > groups.
> >
> > One open question is whether the binary equality operator needs a separate
> > operator class or not. If an opclass cares only about the binary equality, its
> > hash function(s) can be a lot simpler.
>
> Hmm. How about instead adding another regproc field to pg_type which
> stores the OID of a function that tests binary equality for that
> datatype? If that happens to be equal to the OID you got from the
> opclass, then you're all set.

I suppose you mean pg_operator, not pg_type. What I don't like about this is
that the new field would only be useful for very little fraction of
operators.

On the other hand, the drawback of an additional operator classes is that we'd
have to modify almost all the existing operator classes for the hash AM. (The
absence of the new strategy number in an operator class cannot mean that the
existing equality operator can be used to compare binary values too, because
thus we can't guarantee correct behavior of the already existing user-defined
operator classes.)

--
Antonin Houska
Cybertec Schönig & Schönig GmbH
Gröhrmühlgasse 26, A-2700 Wiener Neustadt
Web: https://www.cybertec-postgresql.com