Mybatis一对多查询的两种姿势，你值得拥有（收藏就完事了）

2022-04-09 23:55:11

@[toc]
前言

最近碰到了Mybatis一对多查询的场景，在这里总结对比下常见的两种实现方式。

本文以常见的订单表和订单详情表来举例说明；

数据库表准备

订单表 tbl_order

Mybatis一对多查询的两种姿势，你值得拥有（收藏就完事了）

订单详情表 tlb_order_detail

Mybatis一对多查询的两种姿势，你值得拥有（收藏就完事了）

ps: 一个订单关联多个订单详情，通过order_no订单号关联；

 

实例演示

方法一：联合查询ResultMap映射

sql直接关联查询，然后结果集通过resultMap的collection映射

例如 查询订单列表，包括订单详情

Order.java 中新增字段orderDetailList，用于存详情列表

public class Order {
    private Integer id;

    private String orderNo;

    private Date orderTime;

    private Date payTime;

    private String remark;
    /**订单详情*/
    private List<OrderDetail> orderDetailList;
    //省略get、set

OrderMapper.java 新增查询方法

    List<Order> queryOrderList(Map map);

OrderMapper.xml

<resultMap id="BaseResultMap" type="com.chouxiaozi.mybatisdruid.entity.Order" >
    <id column="id" property="id" jdbcType="INTEGER" />
    <result column="order_no" property="orderNo" jdbcType="VARCHAR" />
    <result column="order_time" property="orderTime" jdbcType="TIMESTAMP" />
    <result column="pay_time" property="payTime" jdbcType="TIMESTAMP" />
    <result column="remark" property="remark" jdbcType="VARCHAR" />
    <collection property="orderDetailList" ofType="com.chouxiaozi.mybatisdruid.entity.OrderDetail">
      <id column="d_id" property="id" jdbcType="INTEGER" />
      <result column="d_order_no" property="orderNo" jdbcType="VARCHAR" />
      <result column="good_name" property="goodName" jdbcType="VARCHAR" />
      <result column="good_id" property="goodId" jdbcType="INTEGER" />
      <result column="good_count" property="goodCount" jdbcType="INTEGER" />
    </collection>
  </resultMap>
  
<select id="queryOrderList" resultMap="BaseResultMap">
    SELECT
      o.*, d.id as d_id,d.order_no as d_order_no,d.good_name,d.good_id,d.good_count
    FROM
      tbl_order o
        LEFT JOIN tbl_order_detail d ON d.order_no = o.order_no
    where 1=1
    <if test="orderNo != null and orderNo != ''">
      and o.order_no = #{orderNo}
    </if>
    ORDER BY o.order_time desc
  </select>

查询结果展示

[
  {
    "id": 2,
    "orderNo": "DD000002",
    "orderTime": "2021-05-09 12:25:57",
    "payTime": "2021-05-09 12:25:59",
    "remark": "2号订单",
    "orderDetailList": [
      {
        "id": 5,
        "orderNo": "DD000002",
        "goodName": "耳机",
        "goodId": 5,
        "goodCount": 1
      },
      {
        "id": 4,
        "orderNo": "DD000002",
        "goodName": "手机",
        "goodId": 4,
        "goodCount": 1
      }
    ]
  },
  {
    "id": 1,
    "orderNo": "DD000001",
    "orderTime": "2021-05-09 12:25:37",
    "payTime": "2021-05-09 12:25:41",
    "remark": "1号订单",
    "orderDetailList": [
      {
        "id": 2,
        "orderNo": "DD000001",
        "goodName": "饮料",
        "goodId": 2,
        "goodCount": 2
      },
      {
        "id": 1,
        "orderNo": "DD000001",
        "goodName": "瓜子",
        "goodId": 1,
        "goodCount": 1
      },
      {
        "id": 3,
        "orderNo": "DD000001",
        "goodName": "矿泉水",
        "goodId": 3,
        "goodCount": 2
      }
    ]
  }
]

原理：sql直接关联查询，然后结果集通过resultMap的collection映射，将order_detail表对应的字段映射到orderDetailList字段中。
优点：条件查询方便；无论是订单表还是详情表如果要进行一些条件过滤的话，非常方便，直接写在where中限制就行。
不足：因为是先关联查询，后映射；如果需要进行分页查询的话，这种方式就无法满足。主表2条数据，详情表5条数据，关联之后就是10条，无法得主表进行分页；解决方法，就是先给主表套个子查询limit分页后，然后结果集再跟详情表进行关联查询；

方法二：子查询映射

通过resultMap中collection标签的select属性去执行子查询

还以查询订单列表为例

OrderMapper.java

    List<Order> queryOrderList2(Map map);

OrderMapper.xml

<!--主查询的resultMap-->
  <resultMap id="BaseResultMap2" type="com.chouxiaozi.mybatisdruid.entity.Order" >
    <id column="id" property="id" jdbcType="INTEGER" />
    <result column="order_no" property="orderNo" jdbcType="VARCHAR" />
    <result column="order_time" property="orderTime" jdbcType="TIMESTAMP" />
    <result column="pay_time" property="payTime" jdbcType="TIMESTAMP" />
    <result column="remark" property="remark" jdbcType="VARCHAR" />
    <!--select子查询， column 传给子查询的参数-->
    <collection property="orderDetailList" ofType="com.chouxiaozi.mybatisdruid.entity.OrderDetail"
                select="queryDetail" column="order_no">
    </collection>
  </resultMap>
  <!--主查询的sql-->
  <select id="queryOrderList2" resultMap="BaseResultMap2">
    SELECT
    o.*
    FROM
    tbl_order o
    where 1=1
    <if test="orderNo != null and orderNo != ''">
      and o.order_no = #{orderNo}
    </if>
    ORDER BY o.order_time desc
  </select>
  <!--子查询的resultMap-->
  <resultMap id="detailResuleMap" type="com.chouxiaozi.mybatisdruid.entity.OrderDetail">
    <id column="id" property="id" jdbcType="INTEGER" />
    <result column="order_no" property="orderNo" jdbcType="VARCHAR" />
    <result column="good_name" property="goodName" jdbcType="VARCHAR" />
    <result column="good_id" property="goodId" jdbcType="INTEGER" />
    <result column="good_count" property="goodCount" jdbcType="INTEGER" />
  </resultMap>
  <!--子查询的sql-->
  <select id="queryDetail" resultMap="detailResuleMap">
    SELECT
      *
    FROM
      `tbl_order_detail` where order_no = #{order_no}
  </select>

查询结果同上个例子一样；

原理：通过collection的select方法去调用子查询；所需参数通过column传递；
优点：无论是分页还是普通查询都能满足；主表增加过滤条件也很方便，直接在主查询的sql中增加where条件就行
缺点：子查询不好增加过滤条件；column只能传递主表已有的字段。下面提供解决方式；
ps：column传递多个参数 column=“{prop1=col1,prop2=col2}”

例如：实际场景中，详情表有个状态字段，只展示状态正常的详情，需要过滤详情记录。
本例子没有状态字段，就查询订单列表，详情中不展示瓜子，即详情记录中过滤掉good_id = 1的；

在上个例子基础上修改如下：
调用层传参

    Map map = new HashMap();
    map.put("goodId", 1);
    orderMapper.queryOrderList2(map);

orderMapper.xml中增加传参过滤
Mybatis一对多查询的两种姿势，你值得拥有（收藏就完事了）

展示结果如下：详情中已成功过滤掉瓜子；记住，过滤子查询不会影响主表记录；

[
  {
    "id": 2,
    "orderNo": "DD000002",
    "orderTime": "2021-05-09 12:25:57",
    "payTime": "2021-05-09 12:25:59",
    "remark": "2号订单",
    "orderDetailList": [
      {
        "id": 4,
        "orderNo": "DD000002",
        "goodName": "手机",
        "goodId": 4,
        "goodCount": 1
      },
      {
        "id": 5,
        "orderNo": "DD000002",
        "goodName": "耳机",
        "goodId": 5,
        "goodCount": 1
      }
    ]
  },
  {
    "id": 1,
    "orderNo": "DD000001",
    "orderTime": "2021-05-09 12:25:37",
    "payTime": "2021-05-09 12:25:41",
    "remark": "1号订单",
    "orderDetailList": [
      {
        "id": 2,
        "orderNo": "DD000001",
        "goodName": "饮料",
        "goodId": 2,
        "goodCount": 2
      },
      {
        "id": 3,
        "orderNo": "DD000001",
        "goodName": "矿泉水",
        "goodId": 3,
        "goodCount": 2
      }
    ]
  }
]

总结

方式 联合查询映射 子查询映射
原理 sql查询完成后再通过resultmap映射结果 主表的数据集循环调用子查询
分页 不支持分页查询，主表套子查询也能实现 支持分页
条件过滤 方便条件过滤 传参也能实现，复杂参数例如list不好传递给子查询 ；子查询过滤不影响主表数据
上一篇：云计算业务拉动业绩增长 各巨头纷纷发力抢滩市场


下一篇：Android SearchView + Toolbar 的语音搜索功能