On mån, 2010-06-07 at 21:33 +0300, Peter Eisentraut wrote:
> I have developed a patch that partially implements the "functional
> dependency" feature that allows some columns to be omitted from the
> GROUP BY clause if it can be shown that the columns are functionally
> dependent on the columns in the group by clause and therefore
> guaranteed to be unique per group.
Second version:
I stripped out all checks except the primary key/unique constraint
checks.
Views whose existence depends on one of those constraints get a
dependency recorded. This depends on the patch currently in the commit
fest to record not null constraints in pg_constraint, so that the
dependencies on not-null constraints can be recorded.
I haven't done any caching of index lookups yet. Some testing with
1600-column tables didn't show any effect. I'll test this a little
more.
