原创

Mybatis系列第7篇:各种查询详解

Mybatis系列目标:从入门开始开始掌握一个高级开发所需要的Mybatis技能。

这是mybatis系列第7篇。

主要内容

  • 单表查询3种方式详解

  • 一对一关联查询(4种方式)详解

  • 一对多查询(2种方式)详解

  • 综合案例
  • 总结
  • 建议
  • 源码

建库建表

创建一个db:javacode2018

4张表:

t_user(用户表)

t_goods(商品表)

t_order(订单表)

t_order_detail(订单明细表)

表之间的关系:

t_order和t_user是一对一的关系,一条订单关联一个用户记录

t_order和t_order_detail是一对多关系,每个订单中可能包含多个子订单,每个子订单对应一个商品

DROP DATABASE IF EXISTS `javacode2018`;
CREATE DATABASE `javacode2018`;

USE `javacode2018`;

DROP TABLE IF EXISTS t_user;
CREATE TABLE t_user(
  id int AUTO_INCREMENT PRIMARY KEY COMMENT '用户id',
  name VARCHAR(32) NOT NULL DEFAULT '' COMMENT '用户名'
) COMMENT '用户表';
INSERT INTO t_user VALUES (1,'张学友'),(2,'路人甲Java');

DROP TABLE IF EXISTS t_goods;
CREATE TABLE t_goods(
  id int AUTO_INCREMENT PRIMARY KEY COMMENT '商品id',
  name VARCHAR(32) NOT NULL DEFAULT '' COMMENT '商品名称',
  price DECIMAL(10,2) NOT NULL DEFAULT 0 COMMENT '商品价格'
) COMMENT '商品信息表';
INSERT INTO t_goods VALUES (1,'Mybatis系列',8.88),(2,'maven高手系列',16.66);

DROP TABLE IF EXISTS t_order;
CREATE TABLE t_order(
  id int AUTO_INCREMENT PRIMARY KEY COMMENT '订单id',
  user_id INT NOT NULL DEFAULT 0 COMMENT '用户id,来源于t_user.id',
  create_time BIGINT NOT NULL DEFAULT 0 COMMENT '订单创建时间(时间戳,秒)',
  up_time BIGINT NOT NULL DEFAULT 0 COMMENT '订单最后修改时间(时间戳,秒)'
) COMMENT '订单表';
INSERT INTO t_order VALUES (1,2,unix_timestamp(now()),unix_timestamp(now())),(2,1,unix_timestamp(now()),unix_timestamp(now()));

DROP TABLE IF EXISTS t_order_detail;
CREATE TABLE t_order_detail(
  id int AUTO_INCREMENT PRIMARY KEY COMMENT '订单明细id',
  order_id INT NOT NULL DEFAULT 0 COMMENT '订单id,来源于t_order.id',
  goods_id INT NOT NULL DEFAULT 0 COMMENT '商品id,来源于t_goods.id',
  num INT NOT NULL DEFAULT 0 COMMENT '商品数量',
  total_price DECIMAL(12,2) NOT NULL DEFAULT 0 COMMENT '商品总金额'
) COMMENT '订单表';
INSERT INTO t_order_detail VALUES (1,1,1,2,17.76),(2,1,1,1,16.66),(3,2,1,1,8.88);

select * from t_user;
select * from t_goods;
select * from t_order;
select * from t_order_detail;

单表查询(3种方式)

需求

需要按照订单id查询订单信息。

方式1

创建每个表对应的Model

db中表的字段是采用下划线分割的,model中我们是采用骆驼命名法来命名的,如OrderModel:

package com.javacode2018.chat05.demo1.model;

import lombok.*;

import java.util.List;

@Getter
@Setter
@Builder
@ToString
@NoArgsConstructor
@AllArgsConstructor
public class OrderModel {
    private Integer id;
    private Integer userId;
    private Long createTime;
    private Long upTime;
}

其他几个Model也类似。

Mapper xml

<select id="getById" resultType="com.javacode2018.chat05.demo1.model.OrderModel">
    <![CDATA[
    SELECT a.id,a.user_id as userId,a.create_time createTime,a.up_time upTime FROM t_order a WHERE a.id = #{value}
    ]]>
</select>

注意上面的resultType,标识结果的类型。

Mapper接口方法

OrderModel getById(int id);

mybatis全局配置文件

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE configuration
        PUBLIC "-//mybatis.org//DTD Config 3.0//EN"
        "http://mybatis.org/dtd/mybatis-3-config.dtd">
<configuration>
    <!-- 引入外部jdbc配置 -->
    <properties resource="jdbc.properties"/>
    <!-- 环境配置,可以配置多个环境 -->
    <environments default="demo4">
        <environment id="demo4">
            <!-- 事务管理器工厂配置 -->
            <transactionManager type="JDBC"/>
            <!-- 数据源工厂配置,使用工厂来创建数据源 -->
            <dataSource type="POOLED">
                <property name="driver" value="${jdbc.driver}"/>
                <property name="url" value="${jdbc.url}"/>
                <property name="username" value="${jdbc.username}"/>
                <property name="password" value="${jdbc.password}"/>
            </dataSource>
        </environment>
    </environments>

    <mappers>
        <mapper resource="demo1/mapper/UserMapper.xml" />
        <mapper resource="demo1/mapper/GoodsMapper.xml" />
        <mapper resource="demo1/mapper/OrderMapper.xml" />
        <mapper resource="demo1/mapper/OrderDetailMapper.xml" />
    </mappers>
</configuration>

测试用例

com.javacode2018.chat05.demo1.Demo1Test#getById

@Before
public void before() throws IOException {
    //指定mybatis全局配置文件
    String resource = "demo1/mybatis-config.xml";
    //读取全局配置文件
    InputStream inputStream = Resources.getResourceAsStream(resource);
    //构建SqlSessionFactory对象
    SqlSessionFactory sqlSessionFactory = new SqlSessionFactoryBuilder().build(inputStream);
    this.sqlSessionFactory = sqlSessionFactory;
}

@Test
public void getById() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById(1);
        log.info("{}", orderModel);
    }
}

运行输出

35:59.211 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById - ==>  Preparing: SELECT a.id,a.user_id as userId,a.create_time createTime,a.up_time upTime FROM t_order a WHERE a.id = ? 
35:59.239 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById - ==> Parameters: 1(Integer)
35:59.258 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById - <==      Total: 1
35:59.258 [main] INFO  c.j.chat05.demo1.Demo1Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790)

原理

sql中我们使用了别名,将t_order中的字段转换成了和OrderModel中字段一样的名称,最后mybatis内部会通过反射,将查询结果按照名称到OrderModel中查找同名的字段,然后进行赋值。

方式2

若我们项目中表对应的Model中的字段都是采用骆驼命名法,mybatis中可以进行一些配置,可以使表中的字段和对应Model中骆驼命名法的字段进行自动映射。

需要在mybatis全局配置文件中加入下面配置:

<settings>
    <!-- 是否开启自动驼峰命名规则映射,及从xx_yy映射到xxYy -->
    <setting name="mapUnderscoreToCamelCase" value="true"/>
</settings>

Mapper xml

<select id="getById1" resultType="com.javacode2018.chat05.demo1.model.OrderModel">
    <![CDATA[
    SELECT a.id,a.user_id,a.create_time,a.up_time FROM t_order a WHERE a.id = #{value}
    ]]>
</select>

注意上面的sql,我们没有写别名了,由于我们开启了自动骆驼命名映射,所以查询结果会按照下面的关系进行自动映射:

sql对应的字段 OrderModel中的字段
id id
user_id userId
create_time createTime
up_time upTime

Mapper接口

OrderModel getById1(int id);

测试用例

com.javacode2018.chat05.demo1.Demo1Test#getById1

@Test
public void getById1() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById1(1);
        log.info("{}", orderModel);
    }
}

运行输出

59:44.884 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==>  Preparing: SELECT a.id,a.user_id,a.create_time,a.up_time FROM t_order a WHERE a.id = ? 
59:44.917 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==> Parameters: 1(Integer)
59:44.935 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - <==      Total: 1
59:44.935 [main] INFO  c.j.chat05.demo1.Demo1Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790)

输出中可以看出,sql中的字段是下划线的方式,OrderModel中的字段是骆驼命名法,结果也自动装配成功,这个就是开启mapUnderscoreToCamelCase产生的效果。

方式3

mapper xml中有个更强大的元素resultMap,通过这个元素可以定义查询结果的映射关系。

Mapper xml

<resultMap id="orderModelMap2" type="com.javacode2018.chat05.demo1.model.OrderModel">
    <id column="id" property="id" />
    <result column="user_id" property="userId" />
    <result column="create_time" property="createTime" />
    <result column="up_time" property="upTime" />
</resultMap>

<select id="getById2" resultMap="orderModelMap2">
    <![CDATA[
    SELECT a.id,a.user_id,a.create_time,a.up_time FROM t_order a WHERE a.id = #{value}
    ]]>
</select>

上面resultMap有2个元素需要指定:

  • id:resultMap标识
  • type:将结果封装成什么类型,此处我们需要将结果分装为OrderModel

注意上面的select元素,有个resultMap,标识查询结果使用哪个resultMap进行映射,此处我们使用的是orderModelMap2,所以查询结果会按照orderModelMap2关联的resultMap进行映射。

Mapper接口

OrderModel getById2(int id);

测试用例

com.javacode2018.chat05.demo1.Demo1Test#getById2

@Test
public void getById2() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById2(1);
        log.info("{}", orderModel);
    }
}

运行输出

14:12.518 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==>  Preparing: SELECT a.id,a.user_id,a.create_time,a.up_time FROM t_order a WHERE a.id = ? 
14:12.546 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==> Parameters: 1(Integer)
14:12.564 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - <==      Total: 1
14:12.564 [main] INFO  c.j.chat05.demo1.Demo1Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790)

一对一关联查询(4种方式)

需求

通过订单id查询订单的时候,将订单关联的用户信息也返回。

我们修改一下OrderModel代码,内部添加一个UserModel,如下:

package com.javacode2018.chat05.demo2.model;

import lombok.*;

import java.util.List;

@Getter
@Setter
@Builder
@ToString
@NoArgsConstructor
@AllArgsConstructor
public class OrderModel {
    private Integer id;
    private Integer userId;
    private Long createTime;
    private Long upTime;

    //下单用户信息
    private UserModel userModel;
}

UserModel内容:

package com.javacode2018.chat05.demo2.model;

import lombok.*;

@Getter
@Setter
@Builder
@ToString
@NoArgsConstructor
@AllArgsConstructor
public class UserModel {
    private Integer id;
    private String name;
}

方式1

Mapper xml

<resultMap id="orderModelMap1" type="com.javacode2018.chat05.demo2.model.OrderModel">
    <id column="id" property="id" />
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <result column="user_id" property="userModel.id"/>
    <result column="name" property="userModel.name"/>
</resultMap>

<select id="getById1" resultMap="orderModelMap1">
    <![CDATA[
    SELECT
        a.id,
        a.user_id,
        a.create_time,
        a.up_time,
        b.name
    FROM
        t_order a,
        t_user b
    WHERE
        a.user_id = b.id
    AND a.id = #{value}
    ]]>
</select>

注意重点在于上面的这两行:

<result column="user_id" property="userModel.id"/>
<result column="name" property="userModel.name"/>

这个地方使用到了级联赋值,多级之间用.进行引用,此处我们只有一级,可以有很多级。

Mapper 接口

OrderModel getById1(int id);

测试用例

com.javacode2018.chat05.demo2.Demo2Test#getById1

@Before
public void before() throws IOException {
    //指定mybatis全局配置文件
    String resource = "demo2/mybatis-config.xml";
    //读取全局配置文件
    InputStream inputStream = Resources.getResourceAsStream(resource);
    //构建SqlSessionFactory对象
    SqlSessionFactory sqlSessionFactory = new SqlSessionFactoryBuilder().build(inputStream);
    this.sqlSessionFactory = sqlSessionFactory;
}

@Test
public void getById1() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById1(1);
        log.info("{}", orderModel);
    }
}

运行输出

24:20.811 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==>  Preparing: SELECT a.id, a.user_id, a.create_time, a.up_time, b.name FROM t_order a, t_user b WHERE a.user_id = b.id AND a.id = ? 
24:20.843 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==> Parameters: 1(Integer)
24:20.861 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - <==      Total: 1
24:20.861 [main] INFO  c.j.chat05.demo2.Demo2Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, userModel=UserModel(id=2, name=路人甲Java))

方式2

这次我们需要使用mapper xml中另外一个元素association,这个元素可以配置关联对象的映射关系,看示例。

Mapper xml

<resultMap id="orderModelMap2" type="com.javacode2018.chat05.demo2.model.OrderModel">
    <id column="id" property="id" />
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <association property="userModel">
        <id column="user_id" property="id"/>
        <result column="name" property="name" />
    </association>
</resultMap>

<select id="getById2" resultMap="orderModelMap2">
    <![CDATA[
    SELECT
        a.id,
        a.user_id,
        a.create_time,
        a.up_time,
        b.name
    FROM
        t_order a,
        t_user b
    WHERE
        a.user_id = b.id
    AND a.id = #{value}
    ]]>
</select>

注意上面下面这部分代码:

<association property="userModel">
    <id column="user_id" property="id"/>
    <result column="name" property="name" />
</association>

注意上面的property属性,这个就是配置sql查询结果和OrderModel.userModel对象的映射关系,将user_iduserModel中的id进行映射,name和userModel中的name进行映射

Mapper接口

OrderModel getById2(int id);

测试用例

@Test
public void getById2() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById2(1);
        log.info("{}", orderModel);
    }
}

运行结果

51:44.896 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==>  Preparing: SELECT a.id, a.user_id, a.create_time, a.up_time, b.name FROM t_order a, t_user b WHERE a.user_id = b.id AND a.id = ? 
51:44.925 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==> Parameters: 1(Integer)
51:44.941 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - <==      Total: 1
51:44.942 [main] INFO  c.j.chat05.demo2.Demo2Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, userModel=UserModel(id=2, name=路人甲Java))

从结果的最后一行可以看出,所有字段的值映射都是ok的。

方式3

先按照订单id查询订单数据,然后在通过订单中user_id去用户表查询用户数据,通过两次查询,组合成目标结果,mybatis已经内置了这种操作,如下。

UserMapper.xml

我们先定义一个通过用户id查询用户信息的select元素,如下

<select id="getById" resultType="com.javacode2018.chat05.demo2.model.UserModel">
    <![CDATA[
    SELECT id,name FROM t_user where id = #{value}
    ]]>
</select>

OrderModel.xml

<resultMap id="orderModelMap3" type="com.javacode2018.chat05.demo2.model.OrderModel">
    <id column="id" property="id" />
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <association property="userModel" select="com.javacode2018.chat05.demo2.mapper.UserMapper.getById" column="user_id" />
</resultMap>

<select id="getById3" resultMap="orderModelMap3">
    <![CDATA[
    SELECT
        a.id,
        a.user_id,
        a.create_time,
        a.up_time
    FROM
        t_order a
    WHERE
        a.id = #{value}
    ]]>
</select>

OrderModel.userModel属性的值来在于另外一个查询,这个查询是通过association元素的select属性指定的,此处使用的是

com.javacode2018.chat05.demo2.mapper.UserMapper.getById

这个查询是有条件的,条件通过associationcolumn进行传递的,此处传递的是getById3查询结果中的user_id字段。

Mapper接口

OrderModel getById3(int id);

测试用例

com.javacode2018.chat05.demo2.Demo2Test#getById3
@Test
public void getById3() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById3(1);
        log.info("{}", orderModel);
    }
}

运行输出

07:12.569 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById3 - ==>  Preparing: SELECT a.id, a.user_id, a.create_time, a.up_time FROM t_order a WHERE a.id = ? 
07:12.600 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById3 - ==> Parameters: 1(Integer)
07:12.619 [main] DEBUG c.j.c.d.mapper.UserMapper.getById - ====>  Preparing: SELECT id,name FROM t_user where id = ? 
07:12.620 [main] DEBUG c.j.c.d.mapper.UserMapper.getById - ====> Parameters: 2(Integer)
07:12.625 [main] DEBUG c.j.c.d.mapper.UserMapper.getById - <====      Total: 1
07:12.625 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById3 - <==      Total: 1
07:12.625 [main] INFO  c.j.chat05.demo2.Demo2Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, userModel=UserModel(id=2, name=路人甲Java))

从输出中可以看出有2次查询,先按照订单id查询订单,然后通过订单记录中用户id去用户表查询用户信息,最终执行了2次查询。

方式4

方式3中给第二个查询传递了一个参数,如果需要给第二个查询传递多个参数怎么办呢?可以这么写

<association property="属性" select="查询对应的select的id" column="{key1=父查询字段1,key2=父查询字段2,key3=父查询字段3}" />

这种相当于给子查询传递了一个map,子查询中 需要用过map的key获取对应的条件,看案例:

OrderMapper.xml

<resultMap id="orderModelMap4" type="com.javacode2018.chat05.demo2.model.OrderModel">
    <id column="id" property="id" />
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <association property="userModel" select="com.javacode2018.chat05.demo2.mapper.UserMapper.getById1" column="{uid1=user_id,uid2=create_time}" />
</resultMap>

<select id="getById4" resultMap="orderModelMap4">
    <![CDATA[
    SELECT
        a.id,
        a.user_id,
        a.create_time,
        a.up_time
    FROM
        t_order a
    WHERE
        a.id = #{value}
    ]]>
</select>

UserMapper.xml

<select id="getById1" resultType="com.javacode2018.chat05.demo2.model.UserModel">
    <![CDATA[
    SELECT id,name FROM t_user where id = #{uid1} and id = #{uid2}
    ]]>
</select>

Mapper接口

OrderModel getById4(int id);

测试用例

com.javacode2018.chat05.demo2.Demo2Test#getById4
@Test
public void getById4() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById4(1);
        log.info("{}", orderModel);
    }
}

运行输出

19:59.881 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById4 - ==>  Preparing: SELECT a.id, a.user_id, a.create_time, a.up_time FROM t_order a WHERE a.id = ? 
19:59.914 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById4 - ==> Parameters: 1(Integer)
19:59.934 [main] DEBUG c.j.c.d.mapper.UserMapper.getById1 - ====>  Preparing: SELECT id,name FROM t_user where id = ? and id = ? 
19:59.934 [main] DEBUG c.j.c.d.mapper.UserMapper.getById1 - ====> Parameters: 2(Integer), 1577947790(Long)
19:59.939 [main] DEBUG c.j.c.d.mapper.UserMapper.getById1 - <====      Total: 0
19:59.939 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById4 - <==      Total: 1
19:59.939 [main] INFO  c.j.chat05.demo2.Demo2Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, userModel=null)

输出中看一下第二个查询的条件,传过来的是第一个查询的user_id和create_time

一对多查询(2种方式)

需求

根据订单id查询出订单信息,并且查询出订单明细列表。

先修改一下OrderModel代码,如下:

package com.javacode2018.chat05.demo3.model;

import lombok.*;

import java.util.List;

@Getter
@Setter
@Builder
@ToString
@NoArgsConstructor
@AllArgsConstructor
public class OrderModel {
    private Integer id;
    private Integer userId;
    private Long createTime;
    private Long upTime;
    //订单详情列表
    private List<OrderDetailModel> orderDetailModelList;
}

OrderModel中添加了一个集合orderDetailModelList用来存放订单详情列表。

方式1

OrderMapper.xml

<resultMap id="orderModelMap1" type="com.javacode2018.chat05.demo3.model.OrderModel">
    <id column="id" property="id"/>
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <collection property="orderDetailModelList" ofType="com.javacode2018.chat05.demo3.model.OrderDetailModel">
        <id column="orderDetailId" property="id"/>
        <result column="order_id" property="orderId"/>
        <result column="goods_id" property="goodsId"/>
        <result column="num" property="num"/>
        <result column="total_price" property="totalPrice"/>
    </collection>
</resultMap>

<select id="getById1" resultMap="orderModelMap1">
    <![CDATA[
    SELECT
        a.id ,
        a.user_id,
        a.create_time,
        a.up_time,
        b.id orderDetailId,
        b.order_id,
        b.goods_id,
        b.num,
        b.total_price
    FROM
        t_order a,
        t_order_detail b
    WHERE
        a.id = b.order_id
        AND a.id = #{value}
    ]]>
</select>

注意上面的getById1中的sql,这个sql中使用到了t_order和t_order_detail连接查询,这个查询会返回多条结果,但是最后结果按照orderModelMap1进行映射,最后只会返回一个OrderModel对象,关键在于collection元素,这个元素用来定义集合中元素的映射关系,有2个属性需要注意:

  • property:对应的属性名称
  • ofType:集合中元素的类型,此处是OrderDetailModel

原理是这样的,注意orderModelMap1中有个

<id column="id" property="id"/>

查询出来的结果会按照这个配置中指定的column进行分组,即按照订单id进行分组,每个订单对应多个订单明细,订单明细会按照collection的配置映射为ofType元素指定的对象。

实际resultMap元素中的id元素可以使用result元素代替,只是用id可以提升性能,mybatis可以通过id元素配置的列的值判断唯一一条记录,如果我们使用result元素,那么判断是否是同一条记录的时候,需要通过所有列去判断了,所以通过id可以提升性能,使用id元素在一对多中可以提升性能,在单表查询中使用id元素还是result元素,性能都是一样的。

Mapper接口

OrderModel getById1(Integer id);

测试用例

com.javacode2018.chat05.demo3.Demo3Test#getById1
@Before
public void before() throws IOException {
    //指定mybatis全局配置文件
    String resource = "demo3/mybatis-config.xml";
    //读取全局配置文件
    InputStream inputStream = Resources.getResourceAsStream(resource);
    //构建SqlSessionFactory对象
    SqlSessionFactory sqlSessionFactory = new SqlSessionFactoryBuilder().build(inputStream);
    this.sqlSessionFactory = sqlSessionFactory;
}

@Test
public void getById1() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        Integer id = 1;
        OrderModel orderModel = mapper.getById1(id);
        log.info("{}", orderModel);
    }
}

运行输出

03:52.092 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==>  Preparing: SELECT a.id , a.user_id, a.create_time, a.up_time, b.id orderDetailId, b.order_id, b.goods_id, b.num, b.total_price FROM t_order a, t_order_detail b WHERE a.id = b.order_id AND a.id = ? 
03:52.124 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - ==> Parameters: 1(Integer)
03:52.148 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById1 - <==      Total: 2
03:52.148 [main] INFO  c.j.chat05.demo3.Demo3Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, orderDetailModelList=[OrderDetailModel(id=1, orderId=1, goodsId=1, num=2, totalPrice=17.76), OrderDetailModel(id=2, orderId=1, goodsId=1, num=1, totalPrice=16.66)])

注意最后一条输出,和期望的结果一致。

方式2

通过2次查询,然后对结果进行分装,先通过订单id查询订单信息,然后通过订单id查询订单明细列表,然后封装结果。mybatis中默认支持这么玩,还是通过collection元素来实现的。

OrderDetailMapper.xml

<select id="getListByOrderId1" resultType="com.javacode2018.chat05.demo3.model.OrderDetailModel" parameterType="int">
    <![CDATA[
    SELECT
        a.id,
        a.order_id AS orderId,
        a.goods_id AS goodsId,
        a.num,
        a.total_price AS totalPrice
    FROM
        t_order_detail a
    WHERE
        a.order_id = #{value}
    ]]>
</select>

OrderMapper.xml

<resultMap id="orderModelMap2" type="com.javacode2018.chat05.demo3.model.OrderModel">
    <id column="id" property="id"/>
    <result column="user_id" property="userId"/>
    <result column="create_time" property="createTime"/>
    <result column="up_time" property="upTime"/>
    <collection property="orderDetailModelList" select="com.javacode2018.chat05.demo3.mapper.OrderDetailMapper.getListByOrderId1" column="id"/>
</resultMap>

<select id="getById2" resultMap="orderModelMap2">
    <![CDATA[
    SELECT
        a.id ,
        a.user_id,
        a.create_time,
        a.up_time
    FROM
        t_order a
    WHERE
        a.id = #{value}
    ]]>
</select>

重点在于下面这句配置:

<collection property="orderDetailModelList" select="com.javacode2018.chat05.demo3.mapper.OrderDetailMapper.getListByOrderId1" column="id"/>

表示orderDetailModelList属性的值通过select属性指定的查询获取,即:

com.javacode2018.chat05.demo3.mapper.OrderDetailMapper.getListByOrderId1

查询参数是通过column属性指定的,此处使用getById2 sql中的id作为条件,即订单id。

Mapper接口

OrderModel getById2(int id);

测试用例

com.javacode2018.chat05.demo3.Demo3Test#getById2

@Test
public void getById2() {
    try (SqlSession sqlSession = this.sqlSessionFactory.openSession(true);) {
        OrderMapper mapper = sqlSession.getMapper(OrderMapper.class);
        OrderModel orderModel = mapper.getById2(1);
        log.info("{}", orderModel);
    }
}

运行输出

10:07.087 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==>  Preparing: SELECT a.id , a.user_id, a.create_time, a.up_time FROM t_order a WHERE a.id = ? 
10:07.117 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - ==> Parameters: 1(Integer)
10:07.135 [main] DEBUG c.j.c.d.m.O.getListByOrderId1 - ====>  Preparing: SELECT a.id, a.order_id AS orderId, a.goods_id AS goodsId, a.num, a.total_price AS totalPrice FROM t_order_detail a WHERE a.order_id = ? 
10:07.136 [main] DEBUG c.j.c.d.m.O.getListByOrderId1 - ====> Parameters: 1(Integer)
10:07.141 [main] DEBUG c.j.c.d.m.O.getListByOrderId1 - <====      Total: 2
10:07.142 [main] DEBUG c.j.c.d.mapper.OrderMapper.getById2 - <==      Total: 1
10:07.142 [main] INFO  c.j.chat05.demo3.Demo3Test - OrderModel(id=1, userId=2, createTime=1577947790, upTime=1577947790, orderDetailModelList=[OrderDetailModel(id=1, orderId=1, goodsId=1, num=2, totalPrice=17.76), OrderDetailModel(id=2, orderId=1, goodsId=1, num=1, totalPrice=16.66)])

输出中有2次查询,先通过订单id查询订单信息,然后通过订单id查询订单明细,mybatis内部对结果进行了组装。

综合案例

入口

com.javacode2018.chat05.demo4.Demo4Test

这个案例中将上面多种查询混合在一起了,有兴趣的可以去看看,加深理解。

总结

  1. mybatis全局配置文件中通过mapUnderscoreToCamelCase可以开启sql中的字段和javabean中的骆驼命名法的字段进行自动映射
  2. 掌握resultMap元素常见的用法
  3. 一对一关联查询使用resultMap->association元素(2种方式)
  4. 一对多查询使用resultMap->collection元素(2种方式)
  5. resultMap中使用id元素主要在复杂的关联查询中可以提升效率,可以通过这个来判断记录的唯一性,如果没有这个,需要通过所有的result相关的列才能判断记录的唯一性

建议

mybatis为我们提供了强大的关联查询,不过个人建议尽量少用,最好是采用单表的方式查询,在程序中通过多次查询,然后自己对结果进行组装。

Model中最好只定义一些和单表字段关联的属性,不要掺杂着其他对象 的引用。

案例代码

链接:https://pan.baidu.com/s/1vt-MAX3oJOu9gyxZAhKkbg 
提取码:i8op

测试用例为代码的入口,下面目录中的所有类:

mybatis-series\chat05\src\test\java\com\javacode2018\chat05

MyBatis系列

  1. MyBatis系列第1篇:MyBatis未出世之前我们那些痛苦的经历
  2. MyBatis系列第2篇:入门篇,带你感受一下mybatis独特的魅力!
  3. MyBatis系列第3篇:Mybatis使用详解(1)
  4. MyBatis系列第4篇:Mybatis使用详解(2)
  5. Mybatis系列第5篇:Mapper接口多种方式传参详解、原理、源码解析
  6. Mybatis系列第6篇:恕我直言,mybatis增删改你未必玩得转!

更多好文章

  1. Java高并发系列(共34篇)
  2. MySql高手系列(共27篇)
  3. Maven高手系列(共10篇)
  4. 聊聊db和缓存一致性常见的实现方式
  5. 接口幂等性这么重要,它是什么?怎么实现?

想了解更多精彩内容请关注我的公众号:路人甲Java

正文到此结束
本文目录