Re: FETCH FIRST clause PERCENT option

From: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
To: Surafel Temesgen <surafel3000(at)gmail(dot)com>
Cc: Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp>, vik(dot)fearing(at)2ndquadrant(dot)com, Mark Dilger <hornschnorter(at)gmail(dot)com>, Andres Freund <andres(at)anarazel(dot)de>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>, andrew(at)tao11(dot)riddles(dot)org(dot)uk
Subject: Re: FETCH FIRST clause PERCENT option
Date: 2019-04-05 15:11:45
Message-ID: 20190405151145.qcxcwxg2bqihzph4@development
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

On Fri, Apr 05, 2019 at 03:14:56PM +0300, Surafel Temesgen wrote:
>On Tue, Mar 26, 2019 at 5:46 PM Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
>wrote:
>
>> On Tue, Mar 26, 2019 at 03:06:52PM +0300, Surafel Temesgen wrote:
>> >On Thu, Feb 28, 2019 at 11:16 PM Tomas Vondra <
>> tomas(dot)vondra(at)2ndquadrant(dot)com>
>> >wrote:
>> >
>> >>
>> >> To give you a (admittedly, somewhat contrived and artificial example):
>> >>
>> >> SELECT * FROM t1 WHERE id IN (
>> >> SELECT id FROM t2 ORDER BY x FETCH FIRST 10 PERCENT ROWS ONLY
>> >> );
>> >>
>> >> Maybe this example is bogus and/or does not really matter in practice. I
>> >> don't know, but I've been unable to convince myself that's the case.
>> >
>> >
>> >does this means we abandon incremental approach? and am not sure of
>> >calculating
>> >percentage after OFFSET clause is acceptable or not
>> >
>>
>> I don't follow. I don't think I've suggested to abandon the incremental
>> approach - I've explained why I think it's what the patch should be doing,
>> and illustrated that with an example.
>>
>
>but it is more complex than the previous approach and it will be more
>complex
>in starting calculating percentage before offset row count. Doesn't
>simplicity prefer?
>

Sure, simplicity has it's value. And often the simplest solution is the
best one. But in But in plenty of cases it's a tradeoff between simplicity
and efficiency, i.e. the simplest solution may perform poorly. I've tried
to explain why I think the incremental solution is the right approach
here - not having to always execute the subplan to completion. I still
haven't heard reasons why that argument is irrelevant (it may be).

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 Antonin Houska 2019-04-05 15:22:42 Re: "WIP: Data at rest encryption" patch and, PostgreSQL 11-beta3
Previous Message Tom Lane 2019-04-05 15:01:42 Re: Failure in contrib test _int on loach