ExplainModifyTarget doesn't work as expected

From: Etsuro Fujita <fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp>
To: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: ExplainModifyTarget doesn't work as expected
Date: 2014-12-22 06:50:52
Message-ID: 5497BF4C.6080302@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

I think ExplainModifyTarget should show the parent of the inheritance
tree in multi-target-table cases, as described there, but noticed that
it doesn't always work like that. Here is an example.

postgres=# create table parent (a int check (a < 0) no inherit);
CREATE TABLE
postgres=# create table child (a int check (a >= 0));
CREATE TABLE
postgres=# alter table child inherit parent;
ALTER TABLE
postgres=# explain update parent set a = a * 2 where a >= 0;
QUERY PLAN
---------------------------------------------------------------
Update on child (cost=0.00..42.00 rows=800 width=10)
-> Seq Scan on child (cost=0.00..42.00 rows=800 width=10)
Filter: (a >= 0)
(3 rows)

IIUC, I think this is because ExplainModifyTarget doesn't take into
account that the parent *can* be excluded by constraint exclusion. So,
I added a field to ModifyTable to record the parent, apart from
resultRelations. (More precisely, the parent in its role as a simple
member of the inheritance tree is recorded so that appending digits to
refname in select_rtable_names_for_explain works as before.) Attached
is a proposed patch for that.

Thanks,

Best regards,
Etsuro Fujita

Attachment Content-Type Size
explain-inherited-updates.patch text/x-patch 5.8 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Paquier 2014-12-22 07:24:52 Commit Fest 2014-12, Status after week 1
Previous Message Tatsuo Ishii 2014-12-22 06:36:36 Re: pgbench -f and vacuum