I have two entities, A and B. I need to easily retrieve entities A, joined with entities B on the condition of equal values of some column (some column from A equal to some column in B). Those columns are not primary or foreign keys, they contain same business data. I need to have access from each instance of A to the collection of B's with the same value of this column. So I model it like this:

This way, I can run queries like "select A join fetch a.bs b where b...."
(Actually, the real relationship here is many-to-many. But when I use @ManyToMany, Hibernate forces me to use join table, which doesnt exist here. So I have to use @OneToMany as workaround).
So far so good. The main problem is: whenever I delete an instance of A, hibernate calls "Update B set column_in_B = null", becuase it thinks the column_in_B is foreign key pointing at primary key in A (and because row in A is deleted, it tries to clean the foreign key in B). BUT the column_in_B IS NOT a foreign key, and can't be modified, because it causes data lost (and this column is NOT NULL anyway in my case, causing data integerity exception to be thrown).

Plese help me with this. How to model such relationships with Hibernate? (I would call it "virtual relationships", or "secondary relationships" or so: as they are not based on foreign keys, they are just some shortcuts which allows for retrieving related objects and quering for them with HQL)

What hibernate version are you using?

How does the mapping at class B look like?

You might be better of doing this join manually with criteria/hql.

3.5

There is no mapping of this relation on B side - this is unidirectional relation (it can't be bidirectional, because A is actually @MappedSuperclass, so it can be a target of relationship). Class B is just simple, plain @Entity with some simple-type attributes.

What do you mean by this? Remove the relationship field "bs" from class A? But how then I get access from each instance of A to the joined instances of B, if there is no field "bs" in A?