I'd like to add a little bit of followup on this issue. The original
blog author posted a new entry the day after I started this thread.
The author indicates that prepare_sql does alleviate the issue.
http://scarydba.wordpress.com/2009/03/11/a-fix-for-nhibernates-parameter-problem/

In that entry the author points to the following:
http://testdrivendevelopment.wordpress.com/2009/03/10/nhibernate-queries-sql-server-execution-plans/

Also related to this topic: The issue came up right out of the gate on
the Herding Code episode 38 podcast that was published on March 13th.
http://herdingcode.com/?p=171


On Mar 10, 4:30 pm, Daniel Auger <[email protected]> wrote:
> So, it looks like the author of the blog post was putting the blame in
> the wrong spot. You are right, the NH team can't fix Sql Server for
> MS.
>
> Fabio, that request could make for an entertaining conversation. ;)
>
> On Mar 10, 3:38 pm, Fabio Maulo <[email protected]> wrote:
>
> > Daniel, if you have time please send the same request to the team are
> > developing your RDBMS and ask they why the system is using a different query
> > plan only because a parameter has a different length... at the end the
> > query-plan should be based on tables and fields involved in the query and
> > not on a parameter value especially when the parameter is not involved in
> > the SELECT clause.
>
> > 2009/3/10 Daniel Auger <[email protected]>
>
> > > Excellent information. Thanks!
>
> > > On Mar 10, 2:07 pm, Fabio Maulo <[email protected]> wrote:
> > > > try to set prepare_sql to true and, btw, it is a specific "problem" of a
> > > > specific RDBMS.
> > > > I know that many people think that NH can solve anything but we can't...
> > > > If an RDBMS has a strange way to paginate result, we can do something 
> > > > but
> > > we
> > > > can't change the RDBMS.
> > > > If WCF has a bug for distributed transaction we can apply, may be, a
> > > > workaroud but the bug is in WCF.
> > > > If a RDBMS change its behavior when you use "OR" clause instead "IN" and
> > > the
> > > > "OR" has better performance then "IN" the real problem is not in NH but
> > > in
> > > > that RDBMS.
>
> > > > We are few, poor and ugly and commercial companies (as Microsoft or
> > > ORACLE)
> > > > has more power to fix their issues.
>
> > > > 2009/3/10 Daniel Auger <[email protected]>
>
> > > > > I hadn't even considered the scariness of that :).
>
> > > > > Are you implying this behavior only happens with primary keys? I was
> > > > > under the impression that any of the varchar parameters could
> > > > > potentially cause another execution plan to be created.
>
> > > > > On Mar 10, 12:35 pm, Gustavo Ringel <[email protected]> wrote:
> > > > > > It is using nvarchar as a primary key, i hope that was what scared
> > > you.
>
> > > > > > Gustavo.
>
> > > > > > On Tue, Mar 10, 2009 at 5:59 PM, Daniel Auger <
> > > [email protected]
> > > > > >wrote:
>
> > > > > > > I came across this today:
>
> > >http://scarydba.wordpress.com/2008/04/29/nhibernate-recompiles-and-ex.
> > > > > ..
>
> > > > > > > I'm wondering how people are dealing with this, or if it is seen 
> > > > > > > as
> > > a
> > > > > > > non issue.
>
> > > > --
> > > > Fabio Maulo
>
> > --
> > Fabio Maulo
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"nhusers" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/nhusers?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to