Re: Fixing the representation of ORDER BY/GROUP BY/DISTINCT

Martijn van Oosterhout <kleptog(at)svana(dot)org> writes:
> On Thu, Jul 31, 2008 at 08:54:49PM -0400, Tom Lane wrote:
>> So while I was fooling with Steve Midgley's problem I got a bit of a bee
>> in my bonnet about the way that the parser emits ORDER BY, GROUP BY,
>> and DISTINCT lists.

> There's an open TODO item in this area: namely that a GROUP BY referring
> to a primary key is equivalent to a GROUP BY involving all the rest of
> the fields. Now, I don't think anyone has proposed a way to support
> that, but I wanted to check we arn't making it harder to do with these
> changes...

No, these changes shouldn't make any difference for that. (Offhand it
seems like that might just be a simple modification in the code that
checks for improper use of ungrouped fields. Not sure though.)

regards, tom lane