Re: Theoretical Basis for SELECT FOR UPDATE
From: x <x_at_not-exists.org>
Date: Fri, 30 Sep 2005 14:55:22 +0300
Message-ID: <dhj93i$8c3$1_at_domitilla.aioe.org>
Date: Fri, 30 Sep 2005 14:55:22 +0300
Message-ID: <dhj93i$8c3$1_at_domitilla.aioe.org>
"Tony Andrews" <andrewst_at_onetel.com> wrote in message
news:1128080541.493388.182640_at_f14g2000cwb.googlegroups.com...
> Marshall Spight wrote:
> I would say that SELECT FOR UPDATE is a valid facility in a DBMS that
> supports multi-statement transactions, even when cursors are not being
> used. It provides a means of locking data that was selected earlier in
> the transaction so that subsequent operations can use the selected data
> without risking violation of ACID properties (e.g. because another user
> deleted that row).
Locking is not part of the logical model. It has no place in a SELECT statement. Received on Fri Sep 30 2005 - 13:55:22 CEST