Re: Isolation levels

From: Lee Fesperman <firstsql_at_ix.netcom.com>
Date: 17 May 2007 11:10:23 -0700
Message-ID: <1179425422.954480.313720_at_n59g2000hsh.googlegroups.com>


Bob Badour wrote:
> Razvan Socol wrote:
> > Please read the following posts by Craig Freedman, member of the SQL
> > Server query execution team:
> > http://blogs.msdn.com/craigfr/archive/2007/04/25/read-committed-isolation-level.aspx
> > http://blogs.msdn.com/craigfr/archive/2007/05/02/query-plans-and-read-committed-isolation-level.aspx
> > http://blogs.msdn.com/craigfr/archive/2007/05/09/repeatable-read-isolation-level.aspx
> >
> > Is the behaviour described in these posts the *correct* behaviour, as
> > defined by ANSI SQL standards, considering the Read Committed and
> > Repetable Read isolation levels, respectively ? Do other DBMS-s behave
> > the same way ? I am aware that the behaviour would be different if we
> > use snapshot isolation, but I'm interested how things *should* work
> > without snapshot isolation.
> >
> > I'd like some responses from people who know really well the ANSI SQL
> > standard and the isolation levels defined in it.

Assuming he is correct about the behavior, SQL Server is compliant with ANSI SQL in the areas he covered. Of course, the 'snapshot' addition is not part of Standard SQL. For more information on transaction isolation, see our tutorial on SQL92, directly at "http:// www.firstsql.com/tutor5.htm#isolation".

His information on SQL Server is incomplete, and I can't tell your about the compliance of other SQL DBMSs except, that is, our product -- FirstSQL/J. FirstSQL/J is fully compliant with the 4 Transaction Isolation Levels of SQL92.

> Thinking you might get a better response from comp.databases, I am
> cross-posting your question there.

Thanks, Bob!

--
Lee Fesperman, FFE Software, Inc. (http://www.firstsql.com)
==============================================================
* The Ultimate DBMS is here!
* FirstSQL/J Object/Relational DBMS  (http://www.firstsql.com)
Received on Thu May 17 2007 - 20:10:23 CEST

Original text of this message