Hi Allen,
This is important information. I have to do more research about the
differences btw. bag and set... :-(
I appreciate your help. Thank you very much!
-- markus.
Allen Gilliland wrote:
k ... tested that out and it causes lots of errors. I have no idea why.
-- Allen
Markus Fuchs wrote:
Hi Allen,
Would you mind experimenting changing hibernate.bag to hibernate.set
and see if it makes any difference?
Thanks!
-- markus.
Allen Gilliland wrote:
I can't answer for this specific situation, but it's certainly
possible that there is no real reason. It's possible that it was a
mistake or it's possible that it doesn't really matter for this
situation and either will work, i dunno.
I would say that in general I think our Hibernate mappings are not
defined to the best of their abilities, so there definitely may be
situations like this where nobody has an exact reason why it is the
way it is :/
-- Allen
Markus Fuchs wrote:
Hi,
I'm wondering, why some collection fields in the pojo classes are
marked
* @hibernate.bag lazy="true" inverse="true" cascade="delete"
and others
* @hibernate.set lazy="true" order-by="name" inverse="true"
cascade="all-delete-orphan"
The two relationships marked as hibernate.bag are
WebSiteData.permissions and UserData.permissions. What special
functionality of the permissions relationships requires the
hibernate.bag annotation? What differentiates hibernate.bag from
hibernate.set? My understanding is that hibernate.set is a special
kind of Set, that can be ordered, etc. hibernate.bag seems to allow
duplicates, is that why it's used for permissions? The application
logic doesn't suggest, that duplicates must be stored. Also,
permissions do have primary keys, and the documentation in
http://anonhibernate.labs.jboss.com/trunk/Hibernate3/src/org/hibernate/mapping/Bag.java
suggests, that the items stored in a Bag don't have ids.
Thanks!
-- markus.
