上两篇文章说了一对一映射，这里说一下多对一 和 一对多的映射情况。
现实中有很多场景需要用到多对一或者一对多，比如上面这两个类图所展现出来的，一般情况下，一个部门会有多名员工，一名员工只在一个部门任职。
多对一关联映射
在上面的场景中，对于Employee来说，它跟Department的关系就是多对一。
PO对象
Employee.java
public class Employee { public int id; public String name; public Department department; //getter、setter }
Department.java
public class Department { public int id; public String name; //getter、setter }
配置多对一关系时，设计po类时，除了写出最基本的属性（比如Employee的id、name），在对应“多”的那个类（比如Employee.java）中添加对应“一”那个类的引用（比如上面的department）。
映射文件
Employee.hbm.xml
<hibernate-mapping package="org.hibernate.test" > <class name="com.danny.hibernate.Employee" table="t_employee"> <id name="id" type="int"> <generator class="native"/> </id> <property name="name"/> <many-to-one name="department" column="departmentid"/> </class> </hibernate-mapping>
Department.hbm.xml
<hibernate-mapping package="org.hibernate.test" > <class name="com.danny.hibernate.Department" table="t_department"> <id name="id"> <generator class="native"/> </id> <property name="name"/> </class> </hibernate-mapping>
映射文件中的内容基本上跟它关联的类中的字段都是对应的。主键配置在<id></id>中，基本字段配置在<property/>中，对其他类的引用配置在<many-to-one/>中。
运行代码执行的建表语句为：
alter table t_employee drop foreign key FKFDCF5A196E78D697 drop table if exists t_department drop table if exists t_employee create table t_department (id integer not null auto_increment, name varchar(255), primary key (id)) create table t_employee (id integer not null auto_increment, name varchar(255), departmentid integer, primary key (id)) alter table t_employee add index FKFDCF5A196E78D697 (departmentid), add constraint FKFDCF5A196E78D697 foreign key (departmentid) references t_department (id)
从建表语句中可以看出来，t_employee表中的外键departmentid与t_department表中的主键相关联。
生成的表结构如下：
下面进行简单测试
插入测试
session.beginTransaction(); Department department=new Department(); department.setName("信息部"); Employee employee1=new Employee(); employee1.setName("小胡"); employee1.setDepartment(department); Employee employee2=new Employee(); employee2.setName("小玉"); employee2.setDepartment(department); session.save(employee1); session.save(employee2); session.getTransaction().commit();
一执行，发现报错了：org.hibernate.TransientObjectException，一看错误就知道，这是因为department还在Transient状态时，session是不能对其操作的。所以可以在事务提交之前先save一下department：
session.beginTransaction(); Department department=new Department(); department.setName("信息部"); Employee employee1=new Employee(); employee1.setName("小胡"); employee1.setDepartment(department); Employee employee2=new Employee(); employee2.setName("小玉"); employee2.setDepartment(department); session.save(department); session.save(employee1); session.save(employee2); session.getTransaction().commit();
这样就可以成功插入了：
还有一种更简单的方法，就是在映射文件Employee.hbm.xml的<many-to-one/>中配置cascade属性，值为"save-update"：
Employee.hbm.xml
<hibernate-mapping package="org.hibernate.test" > <class name="com.danny.hibernate.Employee" table="t_employee"> <id name="id" type="int"> <generator class="native"/> </id> <property name="name"/> <many-to-one name="department" column="departmentid" cascade="save-update"/> </class> </hibernate-mapping>
cascade表示两个对象之间的操作为联动关系，即对一个对象执行了操作之后，对其指定的级联对象也要进行相同的操作。Hibernate文档对cascade的解释为：
＂cascade（级联） (可选): 指明哪些操作会从父对象级联到关联的对象。它的值代表着Hibernate基本操作的名称， persist, merge, delete, save-update, evict, replicate, lock, refresh……＂
查询测试
session.beginTransaction(); Employee employee=(Employee)session.load(Employee.class, 1); System.out.println("employee的name："+employee.getName()); System.out.println("department的name："+employee.getDepartment().getName()); session.getTransaction().commit();
测试结果：
employee的name：小玉 department的name：信息部
一对多关联映射
既然Employee对Department的关系是多对一，那么反之，Department对Employee就是一对多的关系。
所以要在Department的PO类中增加一个Employee对象的集合。这个集合可以是set、list、map甚至array等容器，由于set中的对象不可重复，并且性能更高，所以一般用set。
PO对象
Department.java
public class Department { public int id; public String name; public Set<Employee> employees; //getter、setter }
Employee.java
public class Employee { public int id; public String name; //getter、setter }
映射文件
Employee.hbm.xml
<hibernate-mapping package="org.hibernate.test" > <class name="com.danny.hibernate.Employee" table="t_employee"> <id name="id" type="int"> <generator class="native"/> </id> <property name="name"/> </class> </hibernate-mapping>
Department.hbm.xml
<hibernate-mapping package="org.hibernate.test" > <class name="com.danny.hibernate.Department" table="t_department"> <id name="id"> <generator class="native"/> </id> <property name="name"/> <set name=employees> <key column="departmentid"></key><!-- "多"的一方关联"一"的一方的外键 --> <one-to-many class="com.bjpowernode.hibernate.Employee"/><!-- 一个Department对象对应多个Employee对象 --> </set> </class> </hibernate-mapping>
映射文件Department.hbm.xml中添加了set标签，对应Department类中的集合employees，表示一个Department对象对应多个Employee对象。
插入测试
session.beginTransaction(); Employee employee1=new Employee(); employee1.setName("小玉玉"); Employee employee2=new Employee(); employee2.setName("小洋洋"); Set<Employee> employees=new HashSet<Employee>(); employees.add(employee1); employees.add(employee2); Department department=new Department(); department.setName("信息部"); department.setEmployees(employees); session.save(employee1); session.save(employee2); session.save(department); session.getTransaction().commit();
插入结果：
查询测试
session.beginTransaction(); Department department=(Department)session.load(Department.class,1); System.out.println("department的name："+department.getName()); System.out.println("department的employee有："); for(Employee employee:department.getEmployees()){ System.out.print(" "+employee.getName()); } session.getTransaction().commit();
控制台输出内容为：
Hibernate: select department0_.id as id0_0_, department0_.name as name0_0_ from t_department department0_ where department0_.id=? department的name：信息部 department的employee有：Hibernate: select employees0_.departmentid as departme3_1_, employees0_.id as id1_, employees0_.id as id1_0_, employees0_.name as name1_0_ from t_employee employees0_ where employees0_.departmentid=? 小洋洋 小玉玉
由此可见，一对多的配置中，默认为延迟加载，相当于lazy=”true”。
给映射文件中<set>标签的属性lazy设置为false时，不会延迟加载，即查询Department的时候，会把属于该Department的Employee全部查询出来。控制台输出内容为：
Hibernate: select department0_.id as id0_0_, department0_.name as name0_0_ from t_department department0_ where department0_.id=? Hibernate: select employees0_.departmentid as departme3_1_, employees0_.id as id1_, employees0_.id as id1_0_, employees0_.name as name1_0_ from t_employee employees0_ where employees0_.departmentid=? department的name：信息部 department的employee有： 小洋洋 小玉玉
比较
相同点：映射原理基本一致，建表时，都是在“多”的一端添加外键指向“一”的一端。
区别：维护的关系不同
多对一维护的关系：多指向一的关系，加载“多”的时候可以把“一”也加载出来；
一对多维护的关系：一指向多的关系，加载“一”的时候可以把“多”也加载出来；
