Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands - Mailing list pgsql-hackers

From Fujii Masao
Subject Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands
Date
Msg-id 822e99f5-1b21-5cf2-e38b-9e1161322168@oss.nttdata.com
Whole thread Raw
In response to Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands  (Amul Sul <sulamul@gmail.com>)
Responses Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands  (Amul Sul <sulamul@gmail.com>)
Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands  (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>)
List pgsql-hackers

On 2021/04/10 0:39, Amul Sul wrote:
> On Fri, Apr 9, 2021 at 8:51 PM Bharath Rupireddy
> <bharath.rupireddyforpostgres@gmail.com> wrote:
>>
>> Hi,
>>
>> While checking the ExecuteTruncate code for the FOREIGN TRUNCATE
>> feature, I saw that we filter out the duplicate relations specified in
>> the TRUNCATE command. But before skipping the duplicates, we are just
>> opening the relation, then if it is present in the already seen
>> relids, then closing it and continuing further.
>>
>> I think we can just have the duplicate checking before table_open so
>> that in cases like TRUNCATE foo, foo, foo, foo; we could save costs of
>> table_open and table_close. Attaching a small patch. Thoughts?
>>
>> This is just like what we already do for child tables, see following
>> in ExecuteTruncate:
>>              foreach(child, children)
>>              {
>>                  Oid            childrelid = lfirst_oid(child);
>>
>>                  if (list_member_oid(relids, childrelid))
>>                      continue;
>>
> 
> Well yes, the patch looks pretty much reasonable to be.

LGTM, too. I will commit this patch.
Though that code exists even in older version, I'm not thinking
to back-patch that because it's not a bug.

Regards,

-- 
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION



pgsql-hackers by date:

Previous
From: Amul Sul
Date:
Subject: Re: Avoid unnecessary table open/close for TRUNCATE foo, foo, foo; kind of commands
Next
From: Junduo Dong
Date:
Subject: [GSoC] Metrics and Monitoring for pgagroal