Mybatis分步查询

当我们需要查询的数据分布在多个表中时，我们可以采用联表，但有时候联表语句比较复杂，或者联表后的数据集太大，不但性能上可能会有问题，维护起来也很麻烦。

Mybatis提供了分步查询的功能。分步查询可以将复杂的联表查询分解成多个单表（或多个简单的联表）查询，维护起来更容易，还可以实现延迟加载，节省资源。

Mybatis的分步查询的分步执行流程是：执行上层的查询语句，如果有n条数据数据，m个分步查询，则分n次依次执行这m个分步查询语句，也就是总共执行n × m个分步查询语句；如果某个分步查询语句也有自身的分步查询，则先执行完自身的分步查询再执行下一个分步查询，也就是采用深度递归的方式执行。

---

代码：

1. 分步查询典型用法：

最上层的查询：

    <select id="listClassInfos" resultMap="listClassInfosMap">
        SELECT `id`, `name` FROM `class_info` WHERE NAME LIKE CONCAT('%', #{name}, '%') LIMIT #{pageStart}, #{pageSize}
    select>

    <resultMap id="listClassInfosMap" type="com.example.study.vo.ClassInfoVo">
        <id column="id" property="id" jdbcType="INTEGER" javaType="java.lang.Integer"/>
        <result column="name" property="name" jdbcType="VARCHAR" javaType="java.lang.String"/>
        
        <association property="master" column="{classId = id}"
                     select="com.example.study.mapper.ClassMasterInfoMapper.getMasterByClassId"/>
        
        <collection property="students" column="{classId = id}" select="listStudentsByClassId"/>
        
    resultMap>

    <select id="listStudentsByClassId" resultType="com.example.study.entity.StudentInfoEntity">
        SELECT
        `id` as `id`,
        `name` as `name`,
        `sex` as `sex`,
        `class_id` as `classId`
        FROM `student_info`
        WHERE `class_id` = #{classId}
    select>

ClassMasterInfoMapper.xml中的查询：

    <select id="getMasterByClassId" resultMap="getMasterByClassIdMap">
        SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `class_id` = #{classId}
    select>

    <resultMap id="getMasterByClassIdMap" type="com.example.study.entity.ClassMasterInfoEntity">
        <id column="id" property="id" jdbcType="INTEGER" javaType="java.lang.Integer"/>
        <result column="name" property="name" jdbcType="VARCHAR" javaType="java.lang.String"/>
        <result column="sex" property="sex" jdbcType="VARCHAR" javaType="java.lang.String"/>
        
        <association property="self" column="{selfId = id}" select="self"/>
    resultMap>

    <select id="self" resultType="com.example.study.entity.ClassMasterInfoEntity">
        SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `id` = #{selfId}
    select>

2. 当最上层查询到2条数据时，sql执行顺序和内容如下（从org.apache.ibatis.mapping.MappedStatement.getBoundSql(Object parameterObject方法打断点，执行boundSql.getSql()获得执行的sql内容)：

# 第1个sql：查询最上层的数据，查到2条数据
SELECT `id`, `name` FROM `class_info` WHERE NAME LIKE CONCAT('%', ?, '%') LIMIT ?, ?

# 第2个sql：第1条数据，第1个子查询
SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `class_id` = ?

# 第3个sql：第1条数据，第1个子查询的第1个子查询（只有一个）
SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `id` = ?

# 第4个sql：第1条数据，第2个子查询
SELECT
        `id` as `id`,
        `name` as `name`,
        `sex` as `sex`,
        `class_id` as `classId`
        FROM `student_info`
        WHERE `class_id` = ?

# 第5个sql：第2条数据，第1个子查询
SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `class_id` = ?

# 第6个sql：第2条数据，第1个子查询的第1个子查询（只有一个）
SELECT
        `id` as `id`,
        `class_id` as `classId`,
        `name` as `name`,
        `sex` as `sex`
        FROM `class_master_info`
        WHERE `id` = ?

# 第7个sql：第2条数据，第2个子查询
SELECT
        `id` as `id`,
        `name` as `name`,
        `sex` as `sex`,
        `class_id` as `classId`
        FROM `student_info`
        WHERE `class_id` = ?