From: Scott Ribe <scott_ribe@elevated-dev.com>
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3731.200.110.1.12\))
Subject: regression in PG 15.1
Message-Id: <D86D21A2-20CB-4E28-AF0A-8C2C0139DA97@elevated-dev.com>
Date: Mon, 28 Nov 2022 07:41:24 -0700
To: Pgsql-admin <pgsql-admin@lists.postgresql.org>
Archived-At: 
 <https://www.postgresql.org/message-id/D86D21A2-20CB-4E28-AF0A-8C2C0139DA97%40elevated-dev.com>
Precedence: bulk

Table is partitioned on a column tbl_id, with an increasing "transaction =
id".

  select * from tbl where txid > something and tbl_id =3D someval;

^^^ works as normal, searches the single partition, taking 0.044ms

  select * from tbl where txid > something and tbl_id =3D (select id =
from reftbl where name =3D 'someval');

^^^ fails to exclude partitions, kicks off parallel scans on all 142, =
takes 23,170ms

Queries in question are NOT auto-generated from some ORM, so there is an =
obvious easy workaround. Before trying that, and before digging into =
full table defs here, does anyone have an idea how to nudge the planner =
toward excluding the partitions.

--
Scott Ribe
scott_ribe@elevated-dev.com
https://www.linkedin.com/in/scottribe/