雙向關聯(inverse 的意義)


多對一一對多 中都是單向關聯,也就是其中一方關聯到另一方,而另一方不知道自己被關聯。

如果讓雙方都意識到另一方的存在,這就形成了雙向關聯,在多對一、一對多的例子可以改寫一下,重新設計User類別如下:
  • User.java
package onlyfun.caterpillar;

public class User {
private Long id;
private String name;
private Room room;

public User() {}

public Long getId() {
return id;
}

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

public String getName() {
return name;
}

public void setName(String name) {
this.name = name;
}

public Room getRoom() {
return room;
}

public void setRoom(Room room) {
this.room = room;
}
}

Room類別如下:
  • Room.java
package onlyfun.caterpillar; 

import java.util.Set;

public class Room {
private Long id;
private String address;
private Set users;

public Room() {}

public Long getId() {
return id;
}

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

public String getAddress() {
return address;
}

public void setAddress(String address) {
this.address = address;
}

public Set getUsers() {
return users;
}

public void setUsers(Set users) {
this.users = users;
}

public void addUser(User user) {
users.add(user);
}

public void removeUser(User user) {
users.remove(user);
}
}

如此,User實例可參考至Room實例而維持多對一關係,而Room實例記得User實例而維持一對多關係。

在映射文件方面,可以如下撰寫:
  • User.hbm.xml
<?xml version="1.0" encoding="utf-8"?> 
<!DOCTYPE hibernate-mapping
PUBLIC "-//Hibernate/Hibernate Mapping DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">

<hibernate-mapping>

<class name="onlyfun.caterpillar.User" table="user">
<id name="id" column="id">
<generator class="native"/>
</id>

<property name="name" column="name"/>

<many-to-one name="room"
column="room_id"
class="onlyfun.caterpillar.Room"
cascade="save-update"
outer-join="true"/>
</class>

</hibernate-mapping>

  • Room.hbm.xml
<?xml version="1.0" encoding="utf-8"?> 
<!DOCTYPE hibernate-mapping
PUBLIC "-//Hibernate/Hibernate Mapping DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">

<hibernate-mapping>

<class name="onlyfun.caterpillar.Room" table="room">
<id name="id" column="id">
<generator class="native"/>
</id>

<property name="address" column="address" />

<set name="users" table="user" cascade="save-update">
<key column="room_id"/>
<one-to-many class="onlyfun.caterpillar.User"/>
</set>
</class>

</hibernate-mapping>

映射文件雙方都設定了cascade為save-update,所以您可以用多對一的方式來維持關聯:
User user1 = new User();
user1.setName("bush");
       
User user2 = new User();
user2.setName("caterpillar");

Room room1 = new Room();
room1.setAddress("NTU-M8-419");

user1.setRoom(room1);
user2.setRoom(room1);
       
Session session = HibernateUtil.getSessionFactory().openSession();
Transaction tx = session.beginTransaction();
       
session.save(user1);
session.save(user2);

tx.commit();
session.close();

或是反過來由一對多的方式來維持關聯:
User user1 = new User();
user1.setName("bush");
       
User user2 = new User();
user2.setName("caterpillar");

Room room1 = new Room();
room1.setUsers(new HashSet());
room1.setAddress("NTU-M8-419");
room1.addUser(user1);
room1.addUser(user2);
       
Session session = HibernateUtil.getSessionFactory().openSession();
Transaction tx = session.beginTransaction();
       
session.save(room1);

tx.commit();
session.close();

這邊有個效率議題可以探討,上面的程式片段Hibernate將使用以下的SQL進行儲存:
Hibernate: insert into room (address) values (?)
Hibernate: insert into user (name, room_id) values (?, ?)
Hibernate: insert into user (name, room_id) values (?, ?)
Hibernate: update user set room_id=? where id=?
Hibernate: update user set room_id=? where id=?

上面的程式寫法表示關聯由Room單方面維持,而主控方也是Room,User不知道Room的room_id是多少,所以必須分別儲存Room與 User之後,再更新user的room_id。

在一對多、多對一形成雙向關聯的情況下,可以將關聯維持的控制權交給多的一方,這樣會比較有效率,理由不難理解,就像是在公司中,老闆要記住多個員工的姓 名快,還是每一個員工都記得老闆的姓名快。

所以在一對多、多對一形成雙向關聯的情況下,可以在「一」的一方設定控制權反轉,也就是當儲存「一」的一方時,將關聯維持的控制權交給「多」的一方,以上 面的例子來說,可以設定Room.hbm.xml如下:
  • Room.hbm.xml
<?xml version="1.0" encoding="utf-8"?> 
<!DOCTYPE hibernate-mapping
PUBLIC "-//Hibernate/Hibernate Mapping DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">

<hibernate-mapping>

<class name="onlyfun.caterpillar.Room" table="room">
<id name="id" column="id">
<generator class="native"/>
</id>

<property name="address" column="address" />

<set name="users" table="user" cascade="save-update" inverse="true">
<key column="room_id"/>
<one-to-many class="onlyfun.caterpillar.User"/>
</set>
</class>

</hibernate-mapping>

由於關聯的控制權交給「多」的一方了,所以直接儲存「一」方前,「多」的一方必須意識到「一」的存在,所以程式片段必須改為如下:
User user1 = new User();
user1.setName("bush");
       
User user2 = new User();
user2.setName("caterpillar");

Room room1 = new Room();
room1.setUsers(new HashSet());
room1.setAddress("NTU-M8-419");
room1.addUser(user1);
room1.addUser(user2);

// 多方必須意識到單方的存在
user1.setRoom(room1);

user2.setRoom(room1);
       
Session session = HibernateUtil.getSessionFactory().openSession();
Transaction tx = session.beginTransaction();
       
session.save(room1);

tx.commit();
session.close();

上面的程式片段Hibernate將使用以下的SQL:
Hibernate: insert into room (address) values (?)
Hibernate: insert into user (name, room_id) values (?, ?)
Hibernate: insert into user (name, room_id) values (?, ?)

如果控制權交給另一方了,而另一方沒有意識到對方的存在的話會如何?試著將上面的程式片段中user1.setRoom(room1);與 user2.setRoom(room1);移去,執行之後,您會發現資料庫中room_id會出現null值,這種結果就好比在 多對一 中,您沒有分配給User一個Room,理所當然的,room_id會出現null。