为什么hibernate执行两个查询以急切加载@OneToOne双向关联？

Question

i have entity A that has-a B entity, and B has-a A with @OneToOne bidirectional association.

Now, when i findall A records, hibernate perform two queries with a left outer join on B, something like this:

select a.id, a.id_b, a.field1, b.id, b.field1 from A as a, B as b left outer join b ON b.id=a.id_b;
select a.id, a.id_b, a.field1, b.id, b.field1 from A as a, B as b left outer join b ON b.id=a.id_b WHERE b.id=?

First query load A and B fields and it is ok, but why perform second query to reload A? I think this query load the A content in B, but this A is obviusly the A that contains B... so its already loaded with first query, isn't true?

-- EDIT --

Entity A:

@Entity
public class A implements Serializable{
    // id and other ecc ecc
    @OneToOne
    @JoinColumn(name="id_b")
    B b;
}

Entity B:

@Entity
public class B implements Serializable{
    // id and other ecc ecc
    @OneToOne(mappedBy="b")
    A a;
}

This is the situation, and a findAll on A need two queries... why?

Answer 1

Blow, if A and B share The same primary key column where both entities are joined by using their primary key , you should use @PrimaryKeyJoinColumn instead

@Entity
public class A implements Serializable {

    private MutableInt id = new MutableInt();

    private B b;

    public void setIdAsMutableInt(MutableInt id) {
        this.id = id;
    }

    @Id
    @GeneratedValue
    public Integer getId() {
        return id.intValue();
    }

    public void setId(Integer id) {
        this.id.setValue(id);
    }

    /**
      * Any ToOne annotation, such as @OneToOne and @ManyToOne, is EARGELY loaded, by default
      */
    @OneToOne(fetch=FetchType.LAZY)
    @PrimaryKeyJoinColumn
    @Cascade(CascadeType.SAVE_UPDATE)
    public B getB() {
        return b;
    }

    public void setB(B b) {
        b.setIdAsMutableInt(id);

        this.b = b;
    }

}

And B Notice you do not need mappedBy attribute because of @PrimaryKeyJoinColumn

@Entity
public class B implements Serializable {

    private MutableInt id = new MutableInt();

    private A a;

    public void setIdAsMutableInt(MutableInt id) {
        this.id = id;
    }

    @Id
    public Integer getId() {
        return id.intValue();
    }

    public void setId(Integer id) {
        this.id.setValue(id);
    }

    @OneToOne(fetch=FetchType.LAZY)
    @PrimaryKeyJoinColumn
    public A getA() {
        return a;
    }

    public void setA(A a) {
        this.a = a;
    }

}

Let's Test (You can Test if you want)

A a = new A();
B b = new B();

a.setB(b);

/**
  * b property will be saved because Cascade.SAVE_UPDATE
  */
Serializable id = session.save(a);

b = (B) session
        .createQuery("from B b left join fetch b.a where b.id = :id")
        .setParameter("id", id)
        .list()
        .get(0);

Assert.assertEquals(b.getId(), b.getA().getId());

Notice I use a MutableInt field ( encapsulated by a Integer property) instead of Integer because Integer is a immutable Type as a way Both A and B share The SAME assigned id

But if A and B are joined by using other Than their primary key , you should use @JoinColumn and mappedBy (bi-directional relationship, right) as follows

@Entity
public class A implements Serializable {

    private Integer id;

    private B b;

    @Id
    @GeneratedValue
    public Integer getId() {
        return id;
    }

    public void setId(Integer id) {
        this.id = id;
    }

    /**
      * mappedBy="a" means: Look at "a" field / property at B Entity. If it has any assigned value, join us Through B_ID foreign key column
      */
    @OneToOne(fetch=FetchType.LAZY, mappedBy="a")
    /**
      * Table A has a foreign key column called "B_ID"
      */ 
    @JoinColumn(name="B_ID")
    @Cascade(CascadeType.SAVE_UPDATE)
    public B getB() {
        return b;
    }

    public void setB(B b) {
        this.b = b;
    }

}   

And B

@Entity
public class B implements Serializable {

    private Integer id;

    private A a;

    public void setIdAsMutableInt(MutableInt id) {
        this.id = id;
    }

    @Id
    @GeneratedValue
    public Integer getId() {
        return id;
    }

    public void setId(Integer id) {
        this.id = id;
    }

    @OneToOne(fetch=FetchType.LAZY)
    public A getA() {
        return a;
    }

    public void setA(A a) {
        this.a = a;
    }

}

To test

A a = new A();
B b = new B();

/**
  * Set up both sides
  * Or use some kind of add convenience method
  */ 
a.setB(b);
b.setA(a);

/**
  * b property will be saved because Cascade.SAVE_UPDATE
  */
Serializable id = session.save(a);

b = (B) session
        .createQuery("from B b left join fetch b.a where b.id = :id")
        .setParameter("id", id)
        .list()
        .get(0);

By using The owner side B, you will get Two select statements It occurs because B Table does not contain any foreign key column which points To Table A But by using

"from A a left join fetch ab where a.id = :id"

You will get just one select statement because A knows how To retrieve its joined B by using its B_ID foreign key column

Answer 2

What does your mapping look like exactly?

Do your A and B classes correctly implement hashCode() and equals() so that Hibernate can tell that the A instance pointed to by B is the same instance of the first A ?

Sounds like you are trying to model a bi-directional one-to-one mapping - take a look at the section in the manual on this to see the recommended methods for accomplishing it.