I have an entity where the natural id is made up of 3 columns. Since there are many relationships to and from this entity, I’ve simplified my data model and Hibernate config (allegedly) by using a surrogate key.

In the UI users can create new instances of this entity, add some child objects and, based on the natural id, they can end up creating duplicates.

I’d like to detect these duplicate transient instances on saveOrUpdate() and update the DB accordingly. Unfortunately, since I am using a surrogate key, this won’t work unless I use the 3 column natural id as the primary key.

What do you guys suggest?

I’ve considered creating a event listener, checking for the given entity name, and doing the updates by hand. Preferably I’d like to use an interceptor, as I can limit its scope by applying it in the class performing the saveOrUpdate() call, yet this might require more coding (?) since I have DAOs without transaction demarcation. Using openSession() does not attach to the current thread out-the-box.

My other option (aside from dropping the surrogate key) is to
do the updates by hand in the class that calls saveOrUpdate() prior to making actually making the call.

I’d appreciate any thoughts on this.

Thanks

u can create unique index/alternative key in db

I already have a unique index in the surrogate key. SaveOrUpdate() -merge really, looks at the entities id. Changing that behavior would involve changing the id or creating an event listener or interceptor.

what is your strategy for creating id? if u use some db generation approach it should work