1、一对多
举个例子：一个国家有很多人。一对多
1）表结构
-- 国家country
CREATE TABLE `country` (
`cid` int(5) NOT NULL AUTO_INCREMENT,
`cname` varchar(20) NOT NULL,
PRIMARY KEY (`cid`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8;
-- 人people
CREATE TABLE `people` (
`pid` int(5) NOT NULL AUTO_INCREMENT,
`pname` varchar(20) NOT NULL,
`countryId` int(5) NOT NULL,
PRIMARY KEY (`pid`)
) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=utf8;
在数据库中，哪个表里面有外键，则就是多方
2）实体类
//国家
public class Country {
private Integer cid;
private String cname;
private Set<People> peoples;
//省略get set属性方法
@Override
public String toString() {
return "Country [cid=" + cid + ", cname=" + cname + ", peoples="
+ peoples + "]";
}
}
//人
public class People {
private Integer pid;
private String pname;
private Integer countryId;
//省略get set属性方法
@Override
public String toString() {
return "People [pid=" + pid + ", pname=" + pname + ", countryId="
+ countryId + "]";
}
}
3）Dao接口
public interface ICountryDao {
Country selectById(int id);
Country selectById2(int id);
}
4）mapper.xml文件
<!-- 方法1 多表连接查询方式 -->
<resultMap type="Country" id="countryMapper">
<id column="cid" property="cid"/>
<result column="cname" property="cname"/>
<collection property="peoples" ofType="People">
<id column="pid" property="pid"/>
<result column="pname" property="pname"/>
</collection>
</resultMap>
<select id="selectById" resultMap="countryMapper">
select cid,cname,pid,pname
from country,people
where cid=countryId and cid = #{xxx}
</select>
<!-- 方法2 多表单独查询方式 -->
<select id="selectPeople" resultType="People">
select pid,pname from people where countryId=#{ooo}
</select>
<resultMap type="Country" id="countryMapper2">
<id column="cid" property="cid"/>
<result column="cname" property="cname"/>
<collection property="peoples"
ofType="People"
select="selectPeople"
column="cid" />
</resultMap>
<select id="selectById2" resultMap="countryMapper2">
select cid,cname from country where cid = #{xxx}
</select>
方式一是将多张表先进行连接，连为一张表后进行查询。其查询本质是一张表。也只有一个select
<collection />是集合的意思，即有多个对象。
property：指定关联属性，即Country类中的集合属性
ofType：集合属性的泛型类型
方式二是将主表的查询结果联合其他表的查询结果，封装成一个对象。主表的查询结果中的数据，作为其他表查询的条件。
这多个查询是可以跨越多个映射文件的，即可以跨越多个namespace的。使用时，添加上其所在的namespace即可
关联属性<collection />的数据来源于另一个查询selectPeople，该查询<selectPeople />的动态参数countryId=#{ooo}的值则来自于查询<selectById2 />的查询结果字段cid
5）测试输出
方式一：
[DEBUG] ==> Preparing: select cid,cname,pid,pname from country,people where cid=countryId and cid = ?
[DEBUG] ==> Parameters: 3(Integer)
[TRACE] <== Columns: cid, cname, pid, pname
[TRACE] <== Row: 3, 日本, 1, 新垣结衣
[TRACE] <== Row: 3, 日本, 5, 松岛枫
[DEBUG] <== Total: 2
Country [cid=3, cname=日本, peoples=[People [pid=5, pname=松岛枫, countryId=null], People [pid=1, pname=新垣结衣, countryId=null]]]
方式二
[DEBUG] ==> Preparing: select cid,cname from country where cid = ?
[DEBUG] ==> Parameters: 2(Integer)
[TRACE] <== Columns: cid, cname
[TRACE] <== Row: 2, 美国
[DEBUG] ====> Preparing: select pid,pname from people where countryId=?
[DEBUG] ====> Parameters: 2(Integer)
[TRACE] <==== Columns: pid, pname
[TRACE] <==== Row: 3, 安妮·海瑟薇
[DEBUG] <==== Total: 1
[DEBUG] <== Total: 1
Country [cid=2, cname=美国, peoples=[People [pid=3, pname=安妮·海瑟薇, countryId=null]]]
2、多对一
反过来，多个人对一个国家。每个人只对一个国家。当然不考虑双重国籍的吵架问题。
由于查询多对象时，也是一个一个查的。所以：一对一关联查询的实现方式与多对一的实现方式是相同的。
1）实体类稍有变化：
//国家
public class Country {
private Integer cid;
private String cname;
//省略get set属性方法
@Override
public String toString() {
return "Country [cid=" + cid + ", cname=" + cname + "]";
}
}
//人
public class People {
private Integer pid;
private String pname;
private Integer countryId;
private Country country;
//省略get set属性方法
@Override
public String toString() {
return "People [pid=" + pid + ", pname=" + pname + ", countryId="
+ countryId + ", country=" + country + "]";
}
}
国家没有了Set<People>人成员变量，人有了国家成员变量
2）Dao接口
public interface IPeopleDao {
People selectById(int id);
People selectById2(int id);
}
3）mapper.xml配置文件
<!-- 方法1 多表连接查询方式 -->
<resultMap type="People" id="peopleMapper">
<id column="pid" property="pid"/>
<result column="pname" property="pname"/>
<association property="country" javaType="Country">
<id column="cid" property="cid"/>
<result column="cname" property="cname"/>
</association>
<!-- 用集合的方式尽然可以，集合无非是一个国家的集合 -->
<!--
<collection property="country" ofType="Country">
<id column="cid" property="cid"/>
<result column="cname" property="cname"/>
</collection>
-->
</resultMap>
<select id="selectById" resultMap="peopleMapper">
select pid,pname,cid,cname
from people,country
where pid = #{xxx} and countryId=cid
</select>
<!-- 方法2 多表单独查询方式 -->
<select id="selectCountry" resultType="Country">
select cid,cname from country where cid=#{ooo}
</select>
<resultMap type="People" id="peopleMapper2">
<id column="pid" property="pid"/>
<result column="pname" property="pname"/>
<association property="country"
javaType="Country"
select="selectCountry"
column="countryId" />
</resultMap>
<select id="selectById2" resultMap="peopleMapper2">
select pid,pname,countryId from people where pid = #{xxx}
</select>
方式一<association />标签体现出2个实体对象之间的关联关系，一对一时用。
property：指定关联属性，即People类中的country属性
javaType：关联属性的类型
4）测试输出
方式一：
[DEBUG] ==> Preparing: select pid,pname,cid,cname from people,country where pid = ? and countryId=cid
[DEBUG] ==> Parameters: 5(Integer)
[TRACE] <== Columns: pid, pname, cid, cname
[TRACE] <== Row: 5, 松岛枫, 3, 日本
[DEBUG] <== Total: 1
People [pid=5, pname=松岛枫, countryId=null, country=Country [cid=3, cname=日本]]
方式二：
[DEBUG] ==> Preparing: select pid,pname,countryId from people where pid = ?
[DEBUG] ==> Parameters: 5(Integer)
[TRACE] <== Columns: pid, pname, countryId
[TRACE] <== Row: 5, 松岛枫, 3
[DEBUG] ====> Preparing: select cid,cname from country where cid=?
[DEBUG] ====> Parameters: 3(Integer)
[TRACE] <==== Columns: cid, cname
[TRACE] <==== Row: 3, 日本
[DEBUG] <==== Total: 1
[DEBUG] <== Total: 1
People [pid=5, pname=松岛枫, countryId=null, country=Country [cid=3, cname=日本]]
备注：
Country类中可以不删除Set<People> peoples ; 成员变量。
People类中可以一直有Country country；成员变量。
去掉是为了便于理解一对多、多对一