Andrew Gierth <> writes:
> "Tom" == Tom Lane <> writes:
>  Tom> I think it's a bit too stupid as-is, though.  We don't need to
>  Tom> recalculate for Params in aggdirectargs, do we?

> In theory we would need to.

How come?  Those are only passed to the final function, no?  They
shouldn't affect what is in the hash table.

> But in practice we don't, because we don't
> allow ordered aggs in AGG_HASHED mode anyway.

True, it's moot at the moment.

> We could skip filling in aggParam at all if not in AGG_HASHED mode I
> guess.

Yeah, I'm planning to make it do that.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to