Message-ID: <4561fe9b-e303-dc28-1a3e-1b5aab626395@enterprisedb.com>
Date: Mon, 10 Apr 2023 23:39:07 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.8.0
Subject: Re: Incorrect number of rows inserted into partitioned table
To: =?UTF-8?B?0KDQvtC80LDQvSDQntGB0LjQv9C+0LI=?=
 <osipovromanvladimirovich@gmail.com>, pgsql-bugs@lists.postgresql.org
References: 
 <CAAqk126uv_J-T8MhOvMqmUUD6MPsxFpq=zojDnWeftZokDYyEw@mail.gmail.com>
Content-Language: en-US
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
In-Reply-To: 
 <CAAqk126uv_J-T8MhOvMqmUUD6MPsxFpq=zojDnWeftZokDYyEw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/4561fe9b-e303-dc28-1a3e-1b5aab626395%40enterprisedb.com>
Precedence: bulk

On 4/10/23 20:42, Роман Осипов wrote:
> Good afternoon
> 
> When reloading data from a simple table into a partitioned (through
> inheritance) using a query like:*insert into [new_partition_table]
> select * from [old_table] limit xxxxxxx*;
> 
> There is an insertion not of the amount specified in *limit*,, but a
> little more or less than it.
> 

Which version? I did try this on master, and I can't reproduce it - the
count at the end returns the correct value (1M).

The rule affects the insert status, which looks e.g. like this:

  INSERT 0 611532

instead of 1M, but that's expected because of the rule, I think.

FWIW I wonder why you use rules, it's rather tricky to get that right,
which is why the usual recommendation is not to use this if there's an
alternative way to do stuff (which for partitioning there is).


regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company