Re: path toward faster partition pruning

From: Amit Langote <Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp>
To: David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>
Cc: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: path toward faster partition pruning
Date: 2017-09-15 08:50:17
Message-ID: e02923ea-a117-a6ad-6a3e-ea5e1ba41ece@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

On 2017/09/15 11:16, Amit Langote wrote:
> I will post rebased patches later today, although I think the overall
> design of the patch on the planner side of things is not quite there yet.
> Of course, your and others' feedback is greatly welcome.

Rebased patches attached. Because Dilip complained earlier today about
clauses of the form (const op var) not causing partition-pruning, I've
added code to commute the clause where it is required. Some other
previously mentioned limitations remain -- no handling of OR clauses, no
elimination of redundant clauses for given partitioning column, etc.

A note about 0001: this patch overlaps with
0003-Canonical-partition-scheme.patch from the partitionwise-join patch
series that Ashutosh Bapat posted yesterday [1]. Because I implemented
the planner-portion of this patch based on what 0001 builds, I'm posting
it here. It might actually turn out that we will review and commit
0003-Canonical-partition-scheme.patch on that thread, but meanwhile apply
0001 if you want to play with the later patches. I would certainly like
to review 0003-Canonical-partition-scheme.patch myself, but won't be able
to immediately (see below).

> Also, I must inform to all of those who're looking at this thread that I
> won't be able to respond to emails from tomorrow (9/16, Sat) until 9/23,
> Sat, due to some personal business.

To remind.

Thanks,
Amit

[1]
https://www.postgresql.org/message-id/CAFiTN-skmaqeCVaoAHCBqe2DyfO3f6sgdtEjHWrUgi0kV1yPLQ%40mail.gmail.com

Attachment Content-Type Size
0001-Some-optimizer-data-structures-for-partitioned-rels.patch text/plain 14.0 KB
0002-WIP-planner-side-changes-for-partition-pruning.patch text/plain 18.4 KB
0003-WIP-Interface-changes-for-partition_bound_-cmp-bsear.patch text/plain 10.1 KB
0004-WIP-Implement-get_partitions_for_keys.patch text/plain 7.5 KB
0005-Add-more-tests-for-the-new-partitioning-related-plan.patch text/plain 21.0 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Ashutosh Bapat 2017-09-15 10:11:23 Re: Partition-wise join for join between (declaratively) partitioned tables
Previous Message Kyotaro HORIGUCHI 2017-09-15 08:43:39 Re: [PATCH] Improve geometric types