From: David Conlin <dc345@cantab.net>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-performance@lists.postgresql.org
Subject: Re: Change in CTE treatment in query plans?
Date: Mon, 21 Oct 2019 14:22:01 +0100
Message-ID: <a2e0867f-ef98-471a-c2bf-2184d304f76c@cantab.net> (raw)
In-Reply-To: <19441.1571303072@sss.pgh.pa.us>
References: <560a2cb6-556f-67a7-3676-dba69f2c69e6@cantab.net>
<19441.1571303072@sss.pgh.pa.us>
Hi Tom -
Thanks so much for getting back to me.
I didn't realise that the costs of init/sub plans would be spread across
the call sites - I had (stupidly) assumed that each call site would
include the full cost.
Having taken a couple of days to go back over the problems I was seeing,
you were absolutely right - it was all to do with multiple call sites -
the postgres version was just a red herring.
Thanks for your help & all the best,
Dave
On 17/10/2019 10:04, Tom Lane wrote:
> David Conlin <dc345@cantab.net> writes:
>> Does anyone know if there's been a change in the way values for CTEs are
>> displayed in query plans?
> Offhand I don't recall any such changes, nor does a cursory look
> through explain.c find anything promising.
>
> If you're concerned with a multiply-referenced CTE, one possibility
> for funny results is that the blame for its execution cost could be
> spread across the multiple call sites. The same can happen with
> initplans/subplans. But I'm just guessing; you didn't show any
> concrete examples so it's hard to be definite.
>
> regards, tom lane
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: dc345@cantab.net, tgl@sss.pgh.pa.us, pgsql-performance@lists.postgresql.org
Subject: Re: Change in CTE treatment in query plans?
In-Reply-To: <a2e0867f-ef98-471a-c2bf-2184d304f76c@cantab.net>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox