Re: Improving estimates for TPC-H Q2

From: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
To: Matt Daw <matt(at)mattdaw(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Improving estimates for TPC-H Q2
Date: 2020-05-08 20:04:26
Message-ID: 20200508200426.ct4l5xl4zpnb7hlc@development
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, May 08, 2020 at 07:58:39AM -0400, Matt Daw wrote:
>Hi Tomas, there’s an interesting related paper in the April 2020 PVLDB,
>“Quantifying TPC-H Choke Points and Their Optimizations”:
>http://www.vldb.org/pvldb/vol13/p1206-dreseler.pdf.
>

Thanks.

Seems like an interesting and new paper, although it seems to focus more
on execution than planning, i.e. it discusses execution strategies but
not how to pick the right plan.

E.g. the Q2 might also be rewritten to compute the whole subplan only
once for all ps_partkey values, and then used for lookup (instead of
running it over and over for each ps_partkey value).

But this still depends on us being able to decide between those two
strategies, which relies on good estimates :-(

regards

--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2020-05-08 20:41:54 Back-branch minor release notes are up for review
Previous Message Christoph Berg 2020-05-08 19:45:33 Re: src/test/perl/TestLib.pm: check_pg_config needs /usr/include/postgresql/pg_config.h