Hello,

I use a type of session-per-request-with-detached-objects pattern in my layered application.

I have several situations where Collection's are retrieved from the database and these collections may contain duplicated information (same table row in different Java objects). I can't really help it, the application's domain object graph is not trivial and I can't lock() every transient object I ever retrieved every time I ask Hibernate to open a new Session and perform some operations.

So as there are duplicates in my Collection's, sometimes when I Session.saveOrUpdate() my objects, I get a NonUniqueObjectException.

So I would love Hibernate to simply ignore the duplicate objects and not throw exceptions. Is this possible, is this stupid, anybody met this situation and how did he/she solve it?


Thank you very much,
Adrian.

I don't really get your problem but I would say no.
Using optimistic locking help ?

Hello, thanks for your answer.

I solved the problem by overriding the equals() and hashCode() methods of my domain objects in order for the various HashSet's (my Collections are in fact HashSets) to consider my objects equal() if their id's are equal and not if they are really identical Java objects.

That way two different Java objects with equal id's will never exist in the same HashSet.