If you had a transient object model with a bidirectional parent-child-relationship, could you reparent your child by only setting child.Parent = newParent?
Probably not. You had to ensure within the setter, that the old and new parent's references get updated, too. If you do this, you can simply use cascade=all-delete-orphan and all goes as expected. Don't try to think relational, when you want persistence ignorance. -Markus 2009/9/10 Paladin_za <[email protected]> > > I think NHibernate is a great ORM, but I can't help but feel that in > some cases the product forces me to manage persistence issues it could > have dealt with itself. > > My biggest gripe at the moment is the case of re-parenting. > > For example, I have a parent object with a collection of child objects > and I want to move one of the existing child objects to a different > parent object. > > From a SQL perspective this could be done with the simple update > statement UPDATE ChildTable SET ParentId = DifferentParentId where > ChildId = TheIDOfTheChildToMove. > > But Nhibernate refuses to make this easy. From what I can read on the > web, there are 2 solutions. > > First solution. Change the HasMany mapping on the parent object to use > CascadeAll. This lets re-parenting work fine, but breaks deletes. How > do we typically delete an object in our models, we set the > ChildObject.Parent reference to null and remove the ChildObject from > the Parent's ChildObjects collection. Now NHibernate de-references the > objects, but the data still remains in the database. Outcome = > Developer has to do database programming by calling session.Delete. > > Second solution. Change the HasMany mapping on the parent object to > use CascadeAllDeleteOrphan. Delete works great now. As per the > documentation. amazing. Except, re-parenting breaks. And we get the > stupid object would be resaved exception. To "fix" this the developer > must first "delete" the object from the session and then re-attach it > to the new parent object. > > The above solutions invariably require my domain model to have some > knowledge of the underlying data access implementation. This flies in > the face of persistence ignorance which Nhibernate champions in every > other area. > > Can someone please explain why this is even considered acceptable? > > I can only hope that someone wiser than me has a solution to this. > > > > > --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
