Hibernate Community • View topic - Hibernate and MS sQL server driver problem

View unanswered posts | View active topics

Board index » Hibernate & Java Persistence » Hibernate Users

All times are UTC - 5 hours [ DST ]

Hibernate and MS sQL server driver problem

Page 1 of 1

[ 4 posts ]

Previous topic | Next topic

Author

Message

schnell18

Post subject: Hibernate and MS sQL server driver problem

Posted: Thu Mar 04, 2004 11:16 pm

Newbie

Joined: Wed Aug 27, 2003 5:31 am
Posts: 10

Hi, Hibernate gurus
According to URL:http://www.hibernate.org/74.html#A13
there's a known problem of MS SQL Server driver(Microsoft's official driver) which fails to return correct
batch update number hibernate expected. So hibernate may throw stacktrace like:

Code:

net.sf.hibernate.HibernateException: Batch update row count wrong: 0
   at net.sf.hibernate.impl.BatchingBatcher.doExecuteBatch(BatchingBatcher.java:65)
   at net.sf.hibernate.impl.BatcherImpl.executeBatch(BatcherImpl.java:118)
   at net.sf.hibernate.impl.BatcherImpl.prepareStatement(BatcherImpl.java:55)
   at net.sf.hibernate.impl.BatcherImpl.prepareBatchStatement(BatcherImpl.java:105)
   at net.sf.hibernate.persister.EntityPersister.delete(EntityPersister.java:565)
   at net.sf.hibernate.impl.ScheduledDeletion.execute(ScheduledDeletion.java:29)
   at net.sf.hibernate.impl.SessionImpl.executeAll(SessionImpl.java:2308)
   at net.sf.hibernate.impl.SessionImpl.execute(SessionImpl.java:2266)

We use hibernate 2.1.1. This prolem still crop up.
Is there any patch addressing this issue?
And what's the consequence if so-called // check return codes section in
BatchingBatcher(Line 56-68) is removed.
See $Id: BatchingBatcher.java,v 1.4.2.6 2003/11/27 16:09:59 oneovthafew Exp $

Or can we change the check policy to make expected update count to 0 when underlying
driver is MS SQL Server by introducing a config parameter such as "ignore_batchupdate_bug"?

Regards,
Justin
2004-03-05 11:08

Top

schnell18

Post subject:

Posted: Fri Mar 05, 2004 3:49 am

Newbie

Joined: Wed Aug 27, 2003 5:31 am
Posts: 10

Hi,
After 2 days' scrutiny and hack on Hibernate source code. Finnally, I find the bug has nothing to do with Hibernate or underlying JDBC driver, it was caused by multiple invocations on Session.delete() with the same persistent object. Becuase the object was deleted by first call, the second pass would make batch update count 0, which Hibernate assume it would be 1. And thus the excpetion.

So remember this caveat: Never delete() a persistence object more than once! This may sound stupid, but actually in complicate nested loops it very likely you fail to pay attention to.

Enjoy Hibernate ! ;-)

Regards,
Justin

Top

gavin

Post subject:

Posted: Fri Mar 05, 2004 3:56 am

Hibernate Team

Joined: Tue Aug 26, 2003 12:50 pm
Posts: 5130
Location: Melbourne, Australia

Hmmmmm this raises an interesting point that I had not been aware of before.

This code:

Code:

foo = (Foo) session.load(Foo.class, fooId);
session.delete(foo);
session.delete(foo);

works just fine. This code:

Code:

foo = (Foo) session.load(Foo.class, fooId);
session.delete(foo);
session.flush();
session.delete(foo);

is broken, since flush disassociates deleted instances.

This is Bad. Not sure if anything can be done about it though :(

Damn.

Top

michael

Post subject:

Posted: Fri Mar 05, 2004 4:49 am

Hibernate Team

Joined: Tue Sep 09, 2003 2:10 pm
Posts: 3246
Location: Passau, Germany

Probably it could be catched earlier and produce a nicer error message than "batch update rowcount wrong" ?

Top

Page 1 of 1

[ 4 posts ]

Board index » Hibernate & Java Persistence » Hibernate Users

All times are UTC - 5 hours [ DST ]

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum