Mybatis关联查询与常见问题解决

用户实体类User

public class User implements Serializable {    private Integer id;    private String username;    private String password;    private Date birthday;    private String address;    public User() {    }    public User(String username, String password, Date birthday, String address) {        this.username = username;        this.password = password;        this.birthday = birthday;        this.address = address;    }    get、set、toString方法}

订单实体类Order

public class Order implements Serializable{    private int id;    private Date createTime;    //支付方式 0：支付宝 1：微信    private int channel;    //支付金额    private int payment;    //支付状态 0:待支付 1：未支付 2:已支付    private int status;    public Order(Date createTime, int channel, int payment, int status) {        this.createTime = createTime;        this.channel = channel;        this.payment = payment;        this.status = status;    }    public Order() {    }    get、set、toString方法}

1、单向多对一关联查询

多个订单关联一个客户
在订单实体类中添加用户实体对象，并在数据库表t_order中添加字段user_id作为外键。

Order实体类添加

    //订单关联客户 manyToOne    private User user;    get,set方法，修改toString方法

订单表如下：

用户表如下：

订单sql映射文件Order.xml中添加

<resultMap type="Order" id="order2">        <id column="id" property="id"/>        <result column="create_time" property="createTime"/>        <result column="channel" property="channel"/>        <result column="payment" property="payment"/>        <result column="status" property="status"/>        <!--association ：配置多对一属性（或一对一）            property:order里面的User属性名            javaType:属性类型(这里是用户类型)        -->        <association property="user" javaType="User">        <!-- id:声明主键，表示user_id是关联查询对象的唯一标识-->            <id column="user_id" property="id"/>            <result column="username" property="username"/>            <result column="password" property="password"/>            <result column="birthday" property="birthday"/>            <result column="address" property="address"/>        </association>    </resultMap>

即使字段和属性一样的都要写出来不能省略，不然会不查出来。只会查自己写的要查的字段，其他的字段就算写在了result里面也是不会查出来的。

mapper映射文件OrderMapper中添加方法

public interface OrderMapper {    public Order findByIdWithManyToOne(int id);}

测试：

public class OrderMapperTest {        InputStream inputStream;        SqlSession session;        @Before        public void init() throws Exception{            String resource = "sqlMapConfig.xml";            inputStream = Resources.getResourceAsStream(resource);            SqlSessionFactoryBuilder ssfb = new SqlSessionFactoryBuilder();            SqlSessionFactory sqlSessionFactory = ssfb.build(inputStream);            session = sqlSessionFactory.openSession();        }        @After        public void destroy() throws Exception{            inputStream.close();            session.close();        }        @Test        public void testFindByIdWithManyToOne(){            OrderMapper mapper = session.getMapper(OrderMapper.class);            Order order = mapper.findByIdWithManyToOne(1);            System.out.println(order);        }}

输出：

Order [id=1, createTime=Sat Sep 02 13:21:30 CST 2017, channel=1, payment=999, status=0, user=User [id=1, username=小红, password=123456, birthday=Sat Sep 02 00:00:00 CST 2017, address=安徽合肥, orders=[]]]

如果只查两张表中某一些字段，也可以自定义一个pojo来进行封装。
比如要查订单表的id，支付金额，以及关联用户的id和用户名。

自定义类型

public class BasicDo {    private int order_id;    private int payment;    private int user_id;    private String username;    get、set、toString方法}

映射文件Order.xml中

<select id="findByIdWithPojo" parameterType="int" resultType="BasicDo">        SELECT         o.id order_id,        o.payment,        u.id user_id,        u.username        FROM t_order o         LEFT OUTER JOIN t_user u ON o.user_id=u.id         WHERE o.id=#{id}    </select>

测试：

    @Test        public void testFindByIdWithPojo(){            OrderMapper mapper = session.getMapper(OrderMapper.class);            BasicDo pojo = mapper.findByIdWithPojo(1);            System.out.println(pojo);        }

输出：

BasicDo [order_id=1, payment=999, user_id=1, username=小红]

2、单向一对多查询
一个用户可以有多个订单
现在订单表和用户表如下


用户sql映射文件User.xml中

<!-- 一对多查询 -->    <select id="findByWithOneToMany" parameterType="int" resultMap="user2">        SELECT         *        FROM t_user u LEFT OUTER JOIN t_order o        ON o.user_id = u.id        WHERE u.id=#{id}    </select>    <resultMap type="User" id="user2">        <id column="id" property="id"/>        <result column="username" property="username"/>        <result column="password" property="password"/>        <result column="birthday" property="birthday"/>        <result column="address" property="address"/>        <collection property="orders" ofType="Order">            <id column="id" property="id"/>            <result column="create_time" property="createTime"/>            <result column="channel" property="channel"/>            <result column="payment" property="payment"/>            <result column="status" property="status"/>        </collection>    </resultMap>

mapper接口UserMapper中添加对应方法

public User findByWithOneToMany(int id);

测试：

@Test    public void testFindByWithOneToMany(){        UserMapper mapper = session.getMapper(UserMapper.class);        User user = mapper.findByWithOneToMany(1);        System.out.println(user);    }

输出结果

User [id=1, username=小红, password=123456, birthday=Sat Sep 02 00:00:00 CST 2017, address=安徽合肥, orders=[Order [id=1, createTime=Fri Sep 01 13:15:30 CST 2017, channel=1, payment=888, status=0, user=null]]]

明显错误了，关联的Order的id应该是2而不是1（当成了用户id处理了），而且这里有两个关联的Order，但是也只能列出来第一条。
出现这种错误的原因是两个实体类主键都为id，如果不明确指定那么框架是不能区分的。
解决办法：这里给出个更通用点的，防止弄混那么可以不需要用关联字段user_id，如果主键名一样那么不要用SELECT * ，同时主键需要取别名。
修改如下

<!-- 一对多查询 -->    <select id="findByWithOneToMany" parameterType="int" resultMap="user2">        SELECT         u.id u_id,        u.username,        u.password,        u.birthday,        u.address,        o.id order_id,        o.payment,        o.create_time,        o.channel,        o.status        FROM t_user u LEFT OUTER JOIN t_order o        ON o.user_id = u.id        WHERE u.id=#{id}    </select>    <resultMap type="User" id="user2">        <id column="u_id" property="id"/>        <result column="username" property="username"/>        <result column="password" property="password"/>        <result column="birthday" property="birthday"/>        <result column="address" property="address"/>        <collection property="orders" ofType="Order">            <id column="order_id" property="id"/>            <result column="create_time" property="createTime"/>            <result column="channel" property="channel"/>            <result column="payment" property="payment"/>            <result column="status" property="status"/>        </collection>    </resultMap>

这里讲各自的id都取了别名u_id和order_id至于user_id就不需要去理睬了（虽然可以把user_id作为用户的id但是容易弄混乱）。
测试结果：

User [id=1, username=小红, password=123456, birthday=Sat Sep 02 00:00:00 CST 2017, address=安徽合肥, orders=[Order [id=2, createTime=Fri Sep 01 13:15:30 CST 2017, channel=1, payment=888, status=0, user=null], Order [id=3, createTime=Sat Sep 02 19:45:29 CST 2017, channel=1, payment=777, status=0, user=null]]]

id为2和3的订单都列了出来。