Re: Dead Lock problem with 8.1.3 - Mailing list pgsql-general

From Kai Hessing
Subject Re: Dead Lock problem with 8.1.3
Date
Msg-id 4o2cg3FcmpiiU1@individual.net
Whole thread Raw
In response to Re: Dead Lock problem with 8.1.3  (Kai Hessing <kai.hessing@hobsons.de>)
List pgsql-general
One Addition:

> PostgreSQL 8.1.4 (The one which is taking... a very long time...)
>
> Merge Join  (cost=7751.81..50026810.45 rows=7364 width=4)
>   Merge Cond: ("outer".sid = "inner".sid)
>   ->  Index Scan using stud_pkey on stud s  (cost=7751.81..49994210.01
> rows=56607 width=4)
>         Filter: (NOT (subplan))
>         SubPlan
>           ->  Materialize  (cost=7751.81..8497.94 rows=53613 width=4)
>                 ->  Seq Scan on stud_vera  (cost=0.00..7488.20
> rows=53613 width=4)
>                       Filter: (veraid = 2)
>   ->  Index Scan using stud_vera_sid_veraid_idx on stud_vera v
> (cost=0.00..37646.74 rows=14729 width=4)
>         Index Cond: (veraid = 34)

The interesting thing is that (like I've written in the main post) if I
remove the 'AND veraid=34' it working fine (0.3seconds). The explain for
this working query is:

Merge Join  (cost=45500.73..46741955.20 rows=147288 width=4)
  Merge Cond: ("outer".sid = "inner".sid)
  ->  Index Scan using stud_pkey on stud s  (cost=7734.28..46701101.71
rows=56607 width=4)
        Filter: (NOT (subplan))
        SubPlan
          ->  Materialize  (cost=7734.28..8431.06 rows=50078 width=4)
                ->  Seq Scan on stud_vera  (cost=0.00..7488.20
rows=50078 width=4)
                      Filter: (veraid = 2)
  ->  Sort  (cost=37766.46..38502.90 rows=294576 width=4)
        Sort Key: v.sid
        ->  Seq Scan on stud_vera v  (cost=0.00..6751.76 rows=294576
width=4)

pgsql-general by date:

Previous
From: Kai Hessing
Date:
Subject: Re: Dead Lock problem with 8.1.3
Next
From: Tom Lane
Date:
Subject: Re: Dead Lock problem with 8.1.3