Re: PL/PgSQL, Inheritance, Locks, and Deadlocks - Mailing list pgsql-general

From Greg Stark
Subject Re: PL/PgSQL, Inheritance, Locks, and Deadlocks
Date
Msg-id 87lla7o5vz.fsf@stark.xeocode.com
Whole thread Raw
In response to PL/PgSQL, Inheritance, Locks, and Deadlocks  (Thomas F.O'Connell <tfo@sitening.com>)
Responses Re: PL/PgSQL, Inheritance, Locks, and Deadlocks
List pgsql-general
"Thomas F.O'Connell" <tfo@sitening.com> writes:

>                  UPDATE groups
>                  SET count1 = v_group_count1, count2 = v_group_count2, count3 =

>
> For instance, when run, this stored procedure could try to acquire a lock on
> users2_groups despite not directly referencing it.

Does the users2_groups contain a foreign key reference to the groups table? If
so then if you need to update the groups table regularly you'll want an index
on the referring column. Otherwise in order to check the constraint Postgres
needs to do a sequential scan of the referring table to make sure your update
doesn't break a reference.

I don't know how this plays with locks though.

--
greg

pgsql-general by date:

Previous
From: Greg Stark
Date:
Subject: Re: Does indexing help >= as well as = for integer columns?
Next
From: William Yu
Date:
Subject: Weird PostgreSQL crashes on FC2/FC3 64-bit