Re: Declarative partitioning - another take

From: Amit Langote <Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp>
To: Rajkumar Raghuwanshi <rajkumar(dot)raghuwanshi(at)enterprisedb(dot)com>
Cc: Amit Langote <amitlangote09(at)gmail(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>, pgsql-hackers-owner(at)postgresql(dot)org
Subject: Re: Declarative partitioning - another take
Date: 2017-04-26 02:43:50
Message-ID: a7fc9a40-e742-2619-38dd-7c7ba156e6a1@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

Thanks for testing.

On 2017/04/25 19:03, Rajkumar Raghuwanshi wrote:
> Thanks for looking into it. I have applied fixes and checked for triggers.
> I could see difference in behaviour of statement triggers for INSERT and
> UPDATE, for insert only root partition triggers are getting fired but for
> update root as well as child partition table triggers both getting fired.
> is this expected??

Yes, because I didn't implement anything for the insert case yet. I posed
a question whether to fire partitions' per-statement triggers when
inserting data through the root table.

Robert replied [1] that it would be desirable to not fire partitions'
per-statement triggers if the root table is mentioned in the query; only
fire their per-row triggers if any. It already works that way for
inserts, and applying only 0001 will get you the same for update/delete.
Patch 0002 is to enable firing partition's per-statement triggers even if
the root table is mentioned in the query, but it implemented the same only
for the update/delete cases. If we decide that that's the right thing to
do, then I will implement the same behavior for the insert case too.

Thanks,
Amit

[1]
https://www.postgresql.org/message-id/CA%2BTgmoatBYy8Hyi3cYR1rFrCkD2NM4ZLZcck4QDGvH%3DHddfDwA%40mail.gmail.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Petr Jelinek 2017-04-26 03:35:04 Re: some review comments on logical rep code
Previous Message Amit Langote 2017-04-26 02:34:41 Re: Declarative partitioning - another take