On 04/06/18 09:37, Pavel Stehule wrote: > > > 2018-06-04 9:24 GMT+02:00 Heikki Linnakangas <hlinnaka@iki.fi > <mailto:hlinnaka@iki.fi>>: > > On 04/06/18 09:12, Pavel Stehule wrote: > > It requires introduction of new "safe" functions (& operators). > Immutable > functions are not enough safe. > > CREATE OR REPLACE FUNCTION public.fx() > RETURNS integer > LANGUAGE plpgsql > IMMUTABLE > AS $function$ > BEGIN > RETURN (SELECT count(*) FROM pg_class); > END; > $function$ > > postgres=# SELECT fx(); > ┌─────┐ > │ fx │ > ╞═════╡ > │ 343 │ > └─────┘ > (1 row) > > > That function is incorrectly marked as IMMUTABLE. In that situation, > it's enough that we throw a sane error like "ERROR: no snapshot > available". > > Yes, it is incorrect mark. Unfortunately - this is often workaround for > wrong estimations - so I afraid, in this case, your proposed fix breaks > lot of applications.
I would say such applications are already broken.
I cannot to agree, not in this moment:
1. there is not any workaround, how to force subselect evaluation in planning time - what can be correct for once only evaluated queries.
2. what is not prohibited, is enabled. I agree so this trick is ugly - but I got it from Tom if I remember well maybe more than 10 years ago. Now is too late change it - I think - probably we find more strange features that we hold due compatibility. But this discussion is offtopic for this thread. I am thinking so lot of expressions can be significantly accelerated and I would to check it.
Regards
Pavel
-- Vik Fearing +33 6 46 75 15 36 http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support