Re: Unexpected behaviour of a RAISE statement in an IMMUTABLE function

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Joel Mukuthu <jom(at)upright(dot)co>
Cc: pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: Unexpected behaviour of a RAISE statement in an IMMUTABLE function
Date: 2022-11-23 16:20:26
Message-ID: 3712628.1669220426@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Joel Mukuthu <jom(at)upright(dot)co> writes:
> CREATE FUNCTION raise_exception_immutable(IN err_message text)
> RETURNS void
> LANGUAGE 'plpgsql'
> IMMUTABLE
> AS $BODY$
> BEGIN
> RAISE EXCEPTION
> USING MESSAGE = err_message;
> END;
> $BODY$;

A function with side-effects (like raising an error) isn't
really immutable [1]. We do fudge that a bit, since hardly
anything could be marked immutable if there were a strict
rule about it --- but when the primary point of the function
is to cause that side-effect, you can't fudge it.

> 4. This raises an exception that was surprising to me:

> SELECT raise_exception_immutable('foo') WHERE false;
> -- ERROR: foo
> -- CONTEXT: PL/pgSQL function raise_exception_immutable(text) line 3 at
> RAISE

The allegedly-immutable function is evaluated during constant folding.

> 5. This does not raises an exception, that was also surprising to me:

> SELECT raise_exception_immutable(format('foo')) WHERE false;

format() isn't immutable, only stable; so constant-folding can't
reach the error.

regards, tom lane

[1] https://www.postgresql.org/docs/current/xfunc-volatility.html

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2022-11-23 16:41:48 Re: BUG #17693: Slow performance: Much slower queries on pg_stat_all_tables since 13.4
Previous Message David G. Johnston 2022-11-23 16:14:48 Re: Unexpected behaviour of a RAISE statement in an IMMUTABLE function