Re: BUG #13817: Query planner strange choose while select/count small part of big table - complete

Hi,

On 9.5 beta it still works similar. Did You see my last reply?:

I've done it already. I've also already check rewrited query. No changes.
But i have a idea. There is generated base and there are "normal"
shipment_order_item with 3 - 100 maybe 10000 shipment_order_sub_item but
one has 3 992 102. Could it be a problem?

If yes that is my bad to ask You because I'belive there will not occur in
production and counting could bye match faster.

I think it could be this problem. Additional I checked number of sub items
and there is 6 items and 5 contains 2 sub items and 1 contains 3 992 102
sub items. For me that is only "not normal" data problem and i think it
will be working correct when data is normal.

P.S. I checked all cases on 9.5 and there are no differences.

Regards,
Marcin

2015-12-18 8:09 GMT+01:00 Peter J. Holzer <hjp-pgsql(at)hjp(dot)at>:

> On 2015-12-14 11:45:27 +0000, sienkomarcin(at)gmail(dot)com wrote:
> > The following bug has been logged on the website:
> >
> > Bug reference: 13817
> > Logged by: Marcin_S
> > Email address: sienkomarcin(at)gmail(dot)com
> > PostgreSQL version: 9.4.5
> > Operating system: Windows 7 x64
> > Description:
> >
> > Hi,
> >
> > First sorry for not completed last bug (browser handle it too fast :).
> Here
> > is complete version:
> >
> >
> > I've check todo list but i can't find exact problem i'm reporting. It
> seems
> > like query planner fires not needed sequence scan by all rows in table
> when
> > only a few rows were picked.
>
> Can you try this with version 9.5 (currently in beta)? This looks very
> similar to a problem I reported some time ago and which is fixed in 9.5.
>
> hp
>
> --
> _ | Peter J. Holzer | I want to forget all about both belts and
> |_|_) | | suspenders; instead, I want to buy pants
> | | | hjp(at)hjp(dot)at | that actually fit.
> __/ | http://www.hjp.at/ | -- http://noncombatant.org/
>