Re: planner failure with ProjectSet + aggregation + parallel query

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: planner failure with ProjectSet + aggregation + parallel query
Date: 2018-03-07 20:18:50
Message-ID: CA+TgmoZaQ-+CfzVW7tnGdnZM7k8T88eq92hWiyp=2OWZTrFNpQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Mar 5, 2018 at 10:38 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> While trying to track down a bug today, I found a different bug.
>
> As of 6946280cded903b6f5269fcce105f8ab1d455d33:
>
> rhaas=# create table foo (a int);
> CREATE TABLE
> rhaas=# set min_parallel_table_scan_size = 0;
> SET
> rhaas=# set parallel_setup_cost = 0;
> SET
> rhaas=# set parallel_tuple_cost = 0;
> SET
> rhaas=# select generate_series(1, a) from foo group by a;
> ERROR: ORDER/GROUP BY expression not found in targetlist
>
> Without the SET commands, or without the GROUP BY, or without the SRF,
> it successfully constructs a plan.

I am able to reproduce this on commit
69f4b9c85f168ae006929eec44fc44d569e846b9 (Move targetlist SRF handling
from expression evaluation to new executor node) with the following
modification for a GUC rename:

create table foo (a int);
--set min_parallel_table_scan_size = 0;
set min_parallel_relation_size = 0;
set parallel_setup_cost = 0;
set parallel_tuple_cost = 0;
select generate_series(1, a) from foo group by a;

But on the previous commit I can't reproduce it. So it looks to me
like that's when it got broken.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2018-03-07 20:31:03 Re: Protect syscache from bloating with negative cache entries
Previous Message Daniel Verite 2018-03-07 20:08:43 Re: csv format for psql