Re: SQL Functions and plan time - Mailing list pgsql-general

On Tuesday 08 Jul 2003 7:48 pm, Tom Lane wrote:
> Richard Huxton <dev@archonet.com> writes:
> > Can I ask why, since the plan is constructed at query-time the parameters
> > aren't substitued *before* planning?
>
> Because then the plan couldn't be re-used.  A SQL function may be
> executed many times in a query, so the plan has to be reusable.
> (Or, if you prefer, we have query-level caching of SQL function plans.)

Ah - I see. Fair enough

Thanks Tom.
--
  Richard Huxton