Re: Functional dependencies and GROUP BY

On Sat, Jul 17, 2010 at 11:13, Alex Hunsaker <badalex@gmail.com> wrote:
> On Sat, Jul 17, 2010 at 04:15, Peter Eisentraut <peter_e@gmx.net> wrote:
>> On fre, 2010-07-16 at 22:29 -0600, Alex Hunsaker wrote:
>>> The only corner case I have run into is creating a view with what I
>>> would call an implicit 'not null' constraint.  Demonstration below:
>>>
>>> create table nn (a int4 not null, b int4, unique (a));
>>> select * from nn group by a; -- should this work? I think not?
>>
>> I believe I referred to this upsthread.

Yeah, I went back and reread the thread and um... yep its right where
you posted patch 2.  I think I read it, forgot about it and then it
bubbled up to my subconscious while testing :).

> Anyhow it sounds like I should try it on top of the other patch and
> see if it works.  I assume it might still need some fixups to work
> with that other patch? Or do you expect it to just work?
[ referring to the not null pg_constraint patch ]

Probably no surprise to you, I tried it on top of the not null
pg_constraint patch and it did not work 'out of the box'.  Mainly I
was curious if there was some magic going on that I did not see.  In
any event do you think its worth adding a regression test for this?