| From: | "F(dot) BROUARD / SQLpro" <sqlpro(at)club-internet(dot)fr> |
|---|---|
| To: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: Inoptimal query plan for max() and multicolumn index |
| Date: | 2011-06-21 10:49:02 |
| Message-ID: | 4E00771E.80403@club-internet.fr |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Le 20/06/2011 18:08, Vladimir Kulev a écrit :
>
> Yes, exactly :)
SQL Server does it but PG does not. Expect this for the future....
So try to rewrite the query like this :
select max(timestamp) from sms where number = '5502712'
UNIUON ALL,
select max(timestamp) from sms where number = '5802693'
UNION ALL
select max(timestamp) from sms where number = '5801981'
To see what happen to the query plan !
A +
>
> On Mon, Jun 20, 2011 at 7:41 PM, Kevin Grittner
> <Kevin(dot)Grittner(at)wicourts(dot)gov> wrote:
>> I expect you're hoping for a plan similar to what this gives you?:
>>
>> explain analyze select greatest(
>> (select max(timestamp) from sms where number = '5502712'),
>> (select max(timestamp) from sms where number = '5802693'),
>> (select max(timestamp) from sms where number = '5801981'));
>
--
Frédéric BROUARD - expert SGBDR et SQL - MVP SQL Server - 06 11 86 40 66
Le site sur le langage SQL et les SGBDR : http://sqlpro.developpez.com
Enseignant Arts & Métiers PACA, ISEN Toulon et CESI/EXIA Aix en Provence
Audit, conseil, expertise, formation, modélisation, tuning, optimisation
*********************** http://www.sqlspot.com *************************
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Florian Weimer | 2011-06-21 11:19:43 | Re: Contemplating SSD Hardware RAID |
| Previous Message | Yeb Havinga | 2011-06-21 07:55:32 | Re: Contemplating SSD Hardware RAID |