From: | "Joe Conway" <joseph(dot)conway(at)home(dot)com> |
---|---|
To: | <roypgsqlgen(at)xemaps(dot)com>, "Doug McNaught" <doug(at)wireboard(dot)com> |
Cc: | <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: why no stored procedures? |
Date: | 2001-08-15 16:18:41 |
Message-ID: | 007101c125a5$f348f180$48d210ac@jecw2k1 |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
> roypgsqlgen(at)xemaps(dot)com writes:
>
> > From what I understand, postgresql does not have any of this available
to
> > it. It has procedural languages available to it, but not 'stored
> > procedures'. Functions are fine, but only being able to return one
> > parameter is going to hurt performance since I will have to run more
select
> > statements from the client side to get any other info that my function
might
> > have changed. Plus, from what I read, functions aren't compiled ahead
of
> > time either.
>
> The "functions returning resultsets" problem is definitely being
> looked at. I'm not sure what the status is.
>
> Also, whether functions are compiled and cached depends on the
> procedural language in question. PL/pgSQL definitely does this
> (caches a parse tree of the function). I don't think PL/TCL does,
> but I'm not sure. PL/pgSQL also caches query plans automatically, and
> PL/TCL has support for doing it explicitly.
>
PostgreSQL also supports compiled C functions. This feature has significant
performance advantages over run-of-the-mill stored procedures.
-- Joe
From | Date | Subject | |
---|---|---|---|
Next Message | Pete Leonard | 2001-08-15 16:21:55 | Re: Updating a view |
Previous Message | Ryan C. Bonham | 2001-08-15 15:52:26 | Updating a view |