Re: Interesting case of IMMUTABLE significantly hurting performance

On 4/9/25 22:33, Olleg Samoylov wrote:
>
>
> On 10.04.2025 01:08, Tom Lane wrote:
>> Yeah.  The assumption is that you had a reason for marking the
>> function IMMUTABLE and you want the planner to treat it that way
>> even if it isn't really.  (There are valid use-cases for that, for
>> instance if you want calls to the function to be constant-folded.)
>>             regards, tom lane
>
> Well, to_char(bigint, text) indeed not immutable, because in some
> pattern it uses get information from locale. For instance,'SLDG'
> patterns. But in case of
>
> CREATE OR REPLACE FUNCTION formatted_num_immutable(p_summa bigint)
>  RETURNS text
>  LANGUAGE sql
>  IMMUTABLE STRICT
> RETURN ltrim(to_char(p_summa, '999 999 999 999 999 999 999 999'));
>
> to_char do not use locale information in this pattern. So it is correct
> conclude that to_char is immutable with this pattern and
> formatted_num_immutable too. I did not lie to the planner.

I doubt the code actually looks at the template, it just does the lookup
on pg_proc and sees:

proname | provolatile | prosrc
---------+-------------+---------------------

to_char | s | int8_to_char

>
> So this is looked "strange", immutable function marked as immutable
> function can not be inlined, but exactly the same function marked as
> volatile do.

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com