Re: compute_query_id and pg_stat_statements

st 12. 5. 2021 v 8:10 odesílatel Julien Rouhaud <rjuju123@gmail.com> napsal:

> On Wed, May 12, 2021 at 07:49:13AM +0200, Pavel Stehule wrote:
> >
> > > Ah, I missed that case.  And we are wanting to use pg_stat_statements
> > > with (almost) zero-config?  How about the following behavior?
> > >
> > >
> > Until now, the pg_stat_statements was zero-config. So the change is not
> > user friendly.
>
> Apart from configuring shared_preload_libraries, but agreed.
>
> > The idea so pg_stat_statements requires enabled computed_query_id is not
> > good. There should be dependency only on the queryid column.
>
> I agree that requiring to change compute_query_id when you already added
> pg_stat_statements in shared_preload_libraries isn't good, and the patch I
> sent
> yesterday would fix that.
>

I don't like the idea of implicit force enabling any feature flag, but it
is better than current design. But it doesn't look like a robust solution.

Does it mean that if somebody disables computed_query_id, then
pg_stat_statements will not work?

Why is there the strong dependency between computed_query_id and
pg_stat_statements? Can this dependency be just optional?

Regards

Pavel