Re: [QUESTIONS] locks - Mailing list pgsql-hackers

From Roelof Osinga
Subject Re: [QUESTIONS] locks
Date
Msg-id 35575A59.886FC208@eboa.com
Whole thread Raw
In response to locks  ("Jose' Soares Da Silva" <sferac@bo.nettuno.it>)
List pgsql-hackers
Jose' Soares Da Silva wrote:
>
> Hi all,
>
> After almost one year using PostgreSQL, I'm still discovering things about
> locks, we have little documentation about it:
>
> I know that one can lock a table in the following ways:
>
> * BEGIN statement:
>      If you don't explicit lock a table using LOCK statement, it will be
>      implicit locked only at first UPDATE, INSERT or DELETE operation.
>
> * DECLARE statement:
>      Currently, PostgreSQL doesn't support READ ONLY cursors, once a cursor
>      is declared, other users can only read data referenced by cursor.
>      Write operations to the referenced table like UPDATE, INSERT,
>      DELETE or DROP aren't allowed until the end of transaction.
>
> * LOCK statement:
>      LOCK don't allows read access to locked tables by the other users.
>      If another user try to SELECT a locked table, he must attend
>      until the locked table is released.
>
> Just wondering if there are other ways to lock tables.
> Any replay will be appreciate.                             Thanks, Jose'
>


I believe it to be so that BEGIN locks each table when it first
occurs in a statement. With multiple statements per transaction
this may lead to a deadlock.

Roelof

----------------------------------------------------------------
Home is where the http://eboa.com/ is.
----------------------------------------------------------------

pgsql-hackers by date:

Previous
From: David Hartwig
Date:
Subject: Group/Order by not in target - Was [NEW ODBC DRIVER]
Next
From: ocie@paracel.com
Date:
Subject: Re: [HACKERS] Is there any way to check the status of the Index table.