Re: To commit or not to commit - That is the question?

From: Steve Cosner <stevec_at_zimmer.csufresno.edu>
Date: 1997/09/10
Message-ID: <5v6e82$aq_at_info.csufresno.edu>#1/1

In article <01bcbd3e$23cb9da0$c69d73c1_at_QRPIQ.handc.com>, Ian Quennell <ian.quennell_at_handc.com> wrote:
>Help.
>
>I have a number of SQL*Plus procedures which need to run in sequence. These
>procedures validate and load data in tables and place any validation
>failures in an error table.
>
>However, if, for example the third procedure fails all procedures must be
>rolled back.
>This is fine except that the validation failures are then rolled back as
>well.
>
>Any ideas on how it is possible to keep these failures

The first thing that comes to mind is a two-step process. Do all validation first, recording failures etc. but NOT updating the data. Next, commit the failure records. Then run the same process again, NOT storing validation failures, but instead making your "real" updates.

Steve Cosner Received on Wed Sep 10 1997 - 00:00:00 CEST

This message: [ Message body ]
Next message: Joe Strano: "Re: Change a Property of an Item"
Previous message: Roland Svensson: "Re: To commit or not to commit - That is the question?"
In reply to: Ian Quennell: "To commit or not to commit - That is the question?"
In reply to Ian Quennell: "To commit or not to commit - That is the question?"
Next in thread: Joe Strano: "Re: Change a Property of an Item"

Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ]

Original text of this message