Hello all,

I know this is a silly question and it must have already been discussed here somewhere. My apology but I don't have idea for good keywords to be used for googling. :P

My question is that I don't understand why my call to commit (without saving or updating prior to this commit call) updates my objects in the database. The scenario is that I loaded an object from a database, make some changes to it, and then change my mind not to save it. However, when the commit is called (I have this as a template), it persists the object onto my database.

I wonder if all changes are automatically detected when the commit is called, so what is the point of update or saveOrUpdate functions for the loaded objects.


Thanks

Most of what you want to know is covered in 11.5 and 11.6 in the Hibernate documentation http://docs.jboss.org/hibernate/core/3. ... -modifying. If you change your mind and don't want to store changes to an object, you should call session.evict(object) before the session is committed/flushed.

Hi nordborg,

I don't see any clear explanation there. I still feel it's a bit counter-intuitive. I'm expecting no update to database layer unless save alike functions are called. Otherwise, what is the objective of those functions for ?

Hibernate implements automatic dirty checking so changes are automatically saved when the transaction is committed. This is a fundamental feature of Hibernate (and I suppose of any JPA-compliant implementation). Quote from the documentation:




The save() and update() are used to associate transient or detached objects with a session.

Thanks nordborg,

It's much clearer to me now.

Which may break application layers... Yes it makes things fast and easy but regarding decoupling I would just warn you about sending entities to the UI layer... Which is another story of course :-)