[teiid-dev] Teiid dialect in Hibernate
John Doyle
jdoyle at redhat.com
Wed Mar 24 14:53:04 EDT 2010
We chatted about this over lunch today, and my understanding is that one test will create and drop a table named Foo, and another test will create and drop a table named Foo with different columns. So yeah, same table with different columns.
~jd
----- "Ramesh Reddy" <rareddy at redhat.com> wrote:
> I guess your last statement is the real problem we have in running
> these
> tests against Teiid.
>
> Shelly, do you know if you can generate the DDL per test to avoid the
> issue? Or are you saying in a given Hibernate test uses same table
> with
> multiple different columns?
>
> Ramesh..
>
> On Wed, 2010-03-24 at 13:46 -0400, Shelly McGowan wrote:
> > John,
> >
> > I did some testing with the Hibernate Test Suite using Hibernate
> Branch_3_2_4_SP1_CP. This
> > was just to determine what issues would be exposed with the Dialect.
>
> >
> > The test reports running with the MetaMatrix Dialect and the Oracle
> Dialect can
> > be seen here:
> >
> >
> http://dev39.qa.atl.jboss.com/~smcgowan/hibernate-reports/MetaMatrix/test-reports/
> >
> http://dev39.qa.atl.jboss.com/~smcgowan/hibernate-reports/Oracle/test-reports/
> >
> > I also generated DDL using the MM dialect:
> >
> >
> http://dev39.qa.atl.jboss.com/~smcgowan/hibernate-reports/DDL/DDL-MetaMatrix/
> >
> >
> > After generating the DDL, we populated the databases with the
> tables. We did
> > find, however, that tests reuse the same Table name with a different
> table structure.
> >
> > Shelly
> >
> >
> > ----- Original Message -----
> > From: "John Doyle" <jdoyle at redhat.com>
> > To: "Ramesh Reddy" <rareddy at redhat.com>, "Shelly McGowan"
> <smcgowan at redhat.com>
> > Cc: "teiid-dev" <teiid-dev at lists.jboss.org>
> > Sent: Wednesday, March 24, 2010 10:17:05 AM GMT -05:00 US/Canada
> Eastern
> > Subject: Re: [teiid-dev] Teiid dialect in Hibernate
> >
> > Hey Shelly,
> >
> > You did some work a year or so ago in order to get some AS/Hibernate
> tests working with the Teiid dialect. Do you have any insight or
> suggestions on this issue? How did you workaround the DDL problem?
> >
> > ~jd
> >
> > ----- "Ramesh Reddy" <rareddy at redhat.com> wrote:
> >
> > > I had IRC conversation with Steve Ebersole about adding the Teiid
> > > Dialect to the Hibernate Core in the next upcoming 3.5 release.
> > >
> > > He is not willing to accept any dialects that are not testable
> from
> > > their testing framework. He said Hibernate uses a schema export
> tool
> > > that issues some DDL queries through JDBC interface for their
> testing
> > > to
> > > work.
> > >
> > > Since Teiid driver does not support the DDL currently, issuing
> the
> > > DDL
> > > is not possible. He suggested that we look into ways to
> > > 1) delegate the DDL calls either through the TeiidDriver
> > > (https://jira.jboss.org/jira/browse/TEIID-669)
> > > or
> > > 2) devise a way in their schema export tool to delegate the DDL
> > > class.
> > >
> > > also, since they already have two CR releases, chances of getting
> > > into
> > > 3.5 release seems slim. However, if we can make it that would be
> > > great.
> > >
> > > Any suggestions? volunteers?
> > >
> > > Thanks.
> > >
> > > Ramesh..
> > >
> > > On Tue, 2010-03-23 at 13:05 -0500, Ramesh Reddy wrote:
> > > > I am not dis-agreeing. Just wanted to what our options were.
> > > >
> > > > I do think it is valuable for Teiid be included along with
> > > Hibernate.
> > > > They are about to release 3.5 very soon, we need to see if we
> can
> > > make
> > > > it into 3.5 final release.
> > > >
> > > > So, instead of making this JIRA in the Tools, we need to make
> this
> > > into
> > > > Hibernate by communicating with its lead.
> > > >
> > > >
> > > > Thanks
> > > >
> > > > Ramesh..
> > > >
> > > > On Tue, 2010-03-23 at 11:01 -0400, John Doyle wrote:
> > > > > Ramesh-The drivers are supplied by the DataTools
> > > ConnectionProfiles
> > > > > and Drivers that you have to define before you use Hibernate,
> so
> > > > > drivers are a different issue, and not something that the
> > > hibernate
> > > > > plug-ins have 'built in'. It's an configuration step that's
> > > already
> > > > > extensible in a way that Dialects don't appear to be.
> > > > >
> > > > > I'm personally in favor of including our dialect in Hibernate
> > > because
> > > > > then Tools will just work, there won't be some additional
> step
> > > for
> > > > > Teiid. It seems to me that if Oracle just works out of the
> box
> > > in
> > > > > Tools, then a fellow JBoss project should just work. The
> only
> > > > > downside I can see to including our dialect in Hibernate is
> that
> > > our
> > > > > release cycles don't match up and changes/maintenance would
> have
> > > to be
> > > > > managed across our differing cycles.
> > > >
> > > > _______________________________________________
> > > > teiid-dev mailing list
> > > > teiid-dev at lists.jboss.org
> > > > https://lists.jboss.org/mailman/listinfo/teiid-dev
> > >
> > > _______________________________________________
> > > teiid-dev mailing list
> > > teiid-dev at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/teiid-dev
>
> _______________________________________________
> teiid-dev mailing list
> teiid-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/teiid-dev
More information about the teiid-dev
mailing list