Re: Functionscan estimates

On Fri, Apr 08, 2005 at 04:38:20PM -0600, Michael Fuhr wrote:
> On Fri, Apr 08, 2005 at 03:15:50PM -0700, Josh Berkus wrote:
> >
> > I'm wondering if it might be useful to be able to add estimated selectivity to
> > a function definition for purposes of query estimation. Currently function
> > scans automatically return a flat default 1000 estimated rows. It seems
> > like the DBA ought to be able to ALTER FUNCTION and give it a row estimate
> > for planning purposes.
>
> About a month ago I mentioned that I'd find that useful. In a
> followup, Christopher Kings-Lynne brought up the idea of a GUC
> variable that could give hints about the expected row count.

That seems pretty limited ... what happens if the query contains more
that one SRF?

Maybe issuing some sort of special call to the function (say, with
some boolean in the call info struct) on which it returns planning data;
thus the planner can call the function itself. The hard part would be
figuring out how to do it without breaking backwards compatibility with
functions that don't know how to handle that. (And how to do it in
plpgsql).

--
Alvaro Herrera (<alvherre[(at)]dcc(dot)uchile(dot)cl>)
"La principal característica humana es la tontería"
(Augusto Monterroso)