Re: read uncommited

From: Mark D Powell <Mark.Powell2_at_hp.com>
Date: Fri, 27 Nov 2009 16:45:00 -0800 (PST)
Message-ID: <036a8bc4-bda5-4772-a2f8-6bc0a34e8120_at_m38g2000yqd.googlegroups.com>



On Nov 27, 12:02 pm, "Frank Swarbrick"
<Frank.Swarbr..._at_efirstbank.com> wrote:
> >>> On 11/25/2009 at 12:42 PM, in message <4b0d96b..._at_news.victoria.tc.ca>,
> Malcolm Dew-Jones<yf..._at_vtn1.victoria.tc.ca> wrote:
> > Frank Swarbrick (Frank.Swarbr..._at_efirstbank.com) wrote:
> > : In DB2 you can do something like the following:
>
> > : SELECT *
> > : FROM MYTABLE
> > : WITH UR;
>
> > : If you have, say, a batch application that is inserting a lot of rows
> > in to
> > : a table, with no commits until the end of the job, doing the above
> > query
> > : while the batch update job is still running will allow you to see the
> > : inserted but not yet committed rows.
>
> > : Not something that you would want to do with a user application, but
> > it's
> > : something that as a developer I have found occasionally useful (seeing
> > what
> > : my batch job is "doing" before it's actually done).  
>
> > : Is there any such feature available for Oracle?
>
> > No, Oracle prevents you from seeing data until it's committed.
>
> I did google a bit and thought that Oracle would not allow it, but wanted to
> verify by actually asking the question myself.
> I guess I can understand why the decision was made.  Not sure I agree with
> the reasoning, but...  Probably not a big deal in the end.
> Thanks!
>
> Frank
>
> --
>
> Frank Swarbrick
> Applications Architect - Mainframe Applications Development
> FirstBank Data Corporation - Lakewood, CO  USA
> P: 303-235-1403- Hide quoted text -
>
> - Show quoted text -

Frank, the Oracle default read consistency model is such that a query always returns a time consistent set of data, that is, all rows returned in the query are as the rows existed at the same point in time. The locking scheme is such that writes do not block readers. Hence Oracle has no need to allow dirty reads.

HTH -- Mark D Powell -- Received on Fri Nov 27 2009 - 18:45:00 CST

Original text of this message