Have you tryed adding the mappedBy parameter onto the A field in class B like so
@Entity
class B {
@ManyToMany(cascade=CascadeType.ALL, mappedBy = "b")
private List<A> a;
..
}
Have you tryed adding the mappedBy parameter onto the A field in class B like so
@Entity
class B {
@ManyToMany(cascade=CascadeType.ALL, mappedBy = "b")
private List<A> a;
..
}
Have you specified the inverse join columns?
@Entity
class A {
@ManyToMany(mappedBy="A", cascade=CascadeType.ALL)
private List <B> b;
..
}
@Entity
class B {
@ManyToMany
@JoinTable (
name="A_B",
joinColumns = {@JoinColumn(name="A_ID")},
inverseJoinColumns = {@JoinColumn(name="B_ID")}
)
private List<A> a;
..
}
That's assuming a join table called A_B with columns A_ID and B_ID.
The shortest answer seems to be you cannot and it makes sense. In a bidirectional many-to-many association one side must be master and is used to persist changes to the underlying join table. As JPA will not maintain both side of the association, you could end up with a memory situation that could not be reloaded once stored in the database. Example:
A a1 = new A();
A a2 = new A();
B b = new B();
a1.getB().add(b);
b.getA().add(a2);
If this state could be persisted, you would end up with the following entries in the join table:
a1_id, b_id
a2_id, b_id
But upon loading, how would JPA know that you intended to only let b know about a2 and not a1 ? and what about a2 that should not know about b ?
This is why you have to maintain the bidirectional association yourself (and make the above example impossible to get to, even in memory) and JPA will only persist based on the state of one side of it.
Maybe there is a small mistake in A_M's answer. In my opinion it should be:
@Entity class B { @ManyToMany @JoinTable ( name="A_B", joinColumns = {@JoinColumn(name="B_ID")}, inverseJoinColumns = {@JoinColumn(name="A_ID")} ) private List a; .. }