Re: Using IN with subselect - Mailing list pgsql-general

From Tom Lane
Subject Re: Using IN with subselect
Date
Msg-id 796.1101405421@sss.pgh.pa.us
Whole thread Raw
In response to Re: Using IN with subselect  (Dave Smith <dave.smith@candata.com>)
Responses Re: Using IN with subselect
List pgsql-general
Dave Smith <dave.smith@candata.com> writes:
> Well here is explain. I would guess that it is executed each time ..
> function any different?

> HashAggregate  (cost=288.32..288.32 rows=1 width=32)
>    ->  Hash IN Join  (cost=288.18..288.31 rows=1 width=32)
>          ->  Subquery Scan journal_all  (cost=282.36..282.45 rows=2 width=64)
>          ->  Hash  (cost=5.83..5.83 rows=1 width=13)
>                ->  Index Scan using glmast_index3 on glmast (cost=0.00..5.83 rows=1 width=13)

No ... this plan says to scan glmast once, load the selected rows into
an in-memory hash table, then scan journal_all once and probe the hash
table for matches.  It looks like a pretty decent choice of plan to me.

            regards, tom lane

pgsql-general by date:

Previous
From: Martijn van Oosterhout
Date:
Subject: Re: Using IN with subselect
Next
From: Tom Lane
Date:
Subject: Re: How can I change a cast from explicit only to implicit?