Question

I am working with JPA on a legacy database. The actual semantics are that foo owns the relationship, and that it really only has a single bar, which would require an @ManyToOne annotation.

foo     (id, bar_id)
foo_bar (foo_id, bar_id)
bar     (id)

Anything ending in _id is a foreign key. Also, the columns in foo_bar make up a primary key.

In an ideal world, I would just trash the foo_bar table, but there are other applications that use the same database, which might break if it’s left out. How would I map this relationship completely?

If not, what would be a good way to keep the integrity of this messy relationship?

Answer 1

Map the foo.bar_id column to a ManyToOne relationship. Map the join table to ManyToMany relationship. Don’t provide accessors to the collection of bars in Foo. Make the setBar() clear the collection of bars and add the new one to it.

This way, the ManyToMany association is completely encapsulated inside to ManyToOne association handling.

public class Foo {

    @ManyToOne
    @JoinColumn(name = "bar_id")
    private Bar bar;

    @ManyToMany
    @JoinTable(...)
    private Set<Bar> bars = new HashSet<Bar>();

    // no getter, setter or any other method for the bars collection

    public void setBar(Bar bar) {
        this.bar = bar;
        this.bars.clear();
        if (bar != null) {
            this.bars.add(bar);
        }
    }
}