From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "David Johnston" <polobo(at)yahoo(dot)com> |
Cc: | "'Merlin Moncure'" <mmoncure(at)gmail(dot)com>, "'Adam Mackler'" <adammackler(at)gmail(dot)com>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Views versus user-defined functions: formatting, comments, performance, etc. |
Date: | 2012-08-18 02:00:14 |
Message-ID: | 10411.1345255214@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
"David Johnston" <polobo(at)yahoo(dot)com> writes:
> Trying to answer the previous question this one presented itself: I just
> tried a couple of very simple queries and couldn't get them give me a plan
> that wasn't a "Function Scan". Is it possible that only "scalar" functions
> can be inlined?
> CREATE OR REPLACE FUNCTION three_col_func()
> RETURNS TABLE (col1 varchar, col2 varchar, col3 varchar)
> AS $$
> SELECT '1'::varchar, '2'::varchar, '3'::varchar;
> $$
> LANGUAGE sql
> VOLATILE
> ROWS 1
> ;
No, the problem with that one is that it's marked VOLATILE, which has a
whole set of implications that wouldn't be replicated by an inline
subquery. Try marking it STABLE instead.
There are a pile of other poorly-documented restrictions as well...
one that I notice in a quick look at inline_set_returning_function
is that the function can't be marked STRICT.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Chris Travers | 2012-08-18 02:40:58 | Re: Views versus user-defined functions: formatting, comments, performance, etc. |
Previous Message | David Johnston | 2012-08-18 00:13:17 | Re: Views versus user-defined functions: formatting, comments, performance, etc. |