MySQL 子查询 - 技术栈

子查询指一个查询语句嵌套在另一个查询语句内部的查询，这个特性从MySQL 4.1开始引入

SQL 中子查询的使用大大增强了 SELECT 查询的能力，因为很多时候查询需要从结果集中获取数据，或者需要从同一个表中先计算得出一个数据结果，然后与这个数据结果（可能是某个标量，也可能是某个集合）进行比较

1. 需求分析与问题解决

1.1 实际问题

需求：谁的工资比Abel高

方式一：

可以先查Abel的工资，再找比他的工资高的

sql 复制代码

mysql> use atguigudb;
Database changed
mysql> SELECT
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  last_name = 'Abel';
+----------+
| salary   |
+----------+
| 11000.00 |
+----------+
1 row in set (0.00 sec)

mysql> SELECT
    ->  last_name,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary > 11000;
+-----------+----------+
| last_name | salary   |
+-----------+----------+
| King      | 24000.00 |
| Kochhar   | 17000.00 |
| De Haan   | 17000.00 |
| Greenberg | 12000.00 |
| Russell   | 14000.00 |
| Partners  | 13500.00 |
| Errazuriz | 12000.00 |
| Ozer      | 11500.00 |
| Hartstein | 13000.00 |
| Higgins   | 12000.00 |
+-----------+----------+
10 rows in set (0.00 sec)

方式二：

自连接的方法

sql 复制代码

mysql> SELECT
    ->  e2.last_name,
    ->  e2.salary
    -> FROM
    ->  employees e1,
    ->  employees e2
    -> WHERE
    ->  e1.last_name = 'Abel'
    ->  AND e1.salary < e2.salary;
+-----------+----------+
| last_name | salary   |
+-----------+----------+
| King      | 24000.00 |
| Kochhar   | 17000.00 |
| De Haan   | 17000.00 |
| Greenberg | 12000.00 |
| Russell   | 14000.00 |
| Partners  | 13500.00 |
| Errazuriz | 12000.00 |
| Ozer      | 11500.00 |
| Hartstein | 13000.00 |
| Higgins   | 12000.00 |
+-----------+----------+
10 rows in set (0.00 sec)

方式三：

子查询

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary >
    ->  (
    ->  SELECT
    ->          salary
    ->  FROM
    ->          employees
    ->  WHERE
    ->          last_name = 'Abel'
    ->  );
+-----------+----------+
| last_name | salary   |
+-----------+----------+
| King      | 24000.00 |
| Kochhar   | 17000.00 |
| De Haan   | 17000.00 |
| Greenberg | 12000.00 |
| Russell   | 14000.00 |
| Partners  | 13500.00 |
| Errazuriz | 12000.00 |
| Ozer      | 11500.00 |
| Hartstein | 13000.00 |
| Higgins   | 12000.00 |
+-----------+----------+
10 rows in set (0.00 sec)

称谓规范： 外查询、主查询，内查询、子查询

1.2 子查询的基本使用

子查询的基本语法结构：

子查询（内查询）在主查询之前一次执行完成
子查询的结果被主查询（外查询）使用
注意：
- 子查询要包含在括号内
- 将子查询放在比较条件的右侧 （写在左侧也可以，但是可读性较差）
- 单行操作符对应单行子查询，多行操作符对应多行子查询

1.3 子查询的分类

分类方式1：（按内查询的结果返回一条还是多条记录）

单行子查询
多行子查询

分类方式2：（内查询是否被执行多次）

相关(或关联)子查询

独立运行

子查询不依赖外部查询的任何字段

只执行一次，结果作为一个常量给主查询用

子查询自己能跑完，跟外面没关系
不相关(或非关联)子查询

依赖运行

子查询里引用了外部表的字段

外部表有多少行，子查询就要执行多少次

子查询必须等外面传个值进来才能跑

2. 单行子查询

2.1 单行比较操作符

操作符	含义
=	equal to
>	greater than
>=	greater than or equal to
<	less than
<=	less than or equal to
<>	not equal to

2.2 子查询编写技巧/步骤

从里往外写
从外往里写

2.3 练习

查询工资大于149号员工工资的员工的信息

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  last_name,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary > (
    ->          SELECT
    ->                  salary
    ->          FROM
    ->                  employees
    ->          WHERE
    ->                  employee_id = 149
    ->          );
+-------------+-----------+----------+
| employee_id | last_name | salary   |
+-------------+-----------+----------+
|         100 | King      | 24000.00 |
|         101 | Kochhar   | 17000.00 |
|         102 | De Haan   | 17000.00 |
|         108 | Greenberg | 12000.00 |
|         114 | Raphaely  | 11000.00 |
|         145 | Russell   | 14000.00 |
|         146 | Partners  | 13500.00 |
|         147 | Errazuriz | 12000.00 |
|         148 | Cambrault | 11000.00 |
|         168 | Ozer      | 11500.00 |
|         174 | Abel      | 11000.00 |
|         201 | Hartstein | 13000.00 |
|         205 | Higgins   | 12000.00 |
+-------------+-----------+----------+
13 rows in set (0.00 sec)

返回job_id与141号员工相同，salary比143号员工多的员工姓名，job_id和工资

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  job_id,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary > (
    ->          SELECT
    ->                  salary
    ->          FROM
    ->                  employees
    ->          WHERE
    ->                  employee_id = 143
    ->          )
    ->  AND
    ->  job_id = (
    ->          SELECT
    ->                  job_id
    ->          FROM
    ->                  employees
    ->          WHERE
    ->                  employee_id = 141
    ->          );
+-------------+----------+---------+
| last_name   | job_id   | salary  |
+-------------+----------+---------+
| Nayer       | ST_CLERK | 3200.00 |
| Mikkilineni | ST_CLERK | 2700.00 |
| Bissot      | ST_CLERK | 3300.00 |
| Atkinson    | ST_CLERK | 2800.00 |
| Mallin      | ST_CLERK | 3300.00 |
| Rogers      | ST_CLERK | 2900.00 |
| Ladwig      | ST_CLERK | 3600.00 |
| Stiles      | ST_CLERK | 3200.00 |
| Seo         | ST_CLERK | 2700.00 |
| Rajs        | ST_CLERK | 3500.00 |
| Davies      | ST_CLERK | 3100.00 |
+-------------+----------+---------+
11 rows in set (0.00 sec)

返回公司工资最少的员工的last_name,job_id和salary

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  job_id,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary = (
    ->          SELECT MIN(salary)
    ->          FROM employees
    ->          );
+-----------+----------+---------+
| last_name | job_id   | salary  |
+-----------+----------+---------+
| Olson     | ST_CLERK | 2100.00 |
+-----------+----------+---------+
1 row in set (0.00 sec)

查询与141号或174号员工的manager_id和department_id相同的其他员工的employee_id， manager_id，department_id

方法一：不成对比较

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  manager_id,
    ->  department_id
    -> FROM
    ->  employees
    -> WHERE
    ->  manager_id IN (
    ->          SELECT manager_id
    ->          FROM employees
    ->          WHERE employee_id IN(174,141)
    ->          )
    ->  AND
    ->  department_id IN (
    ->          SELECT department_id
    ->          FROM employees
    ->          WHERE employee_id IN(174,141)
    ->          )
    ->  AND
    ->  employee_id NOT IN(174,141);
+-------------+------------+---------------+
| employee_id | manager_id | department_id |
+-------------+------------+---------------+
|         142 |        124 |            50 |
|         143 |        124 |            50 |
|         144 |        124 |            50 |
|         196 |        124 |            50 |
|         197 |        124 |            50 |
|         198 |        124 |            50 |
|         199 |        124 |            50 |
|         175 |        149 |            80 |
|         176 |        149 |            80 |
|         177 |        149 |            80 |
|         179 |        149 |            80 |
+-------------+------------+---------------+
11 rows in set (0.00 sec)

方法二：成对比较（了解）

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  manager_id,
    ->  department_id
    -> FROM
    ->  employees
    -> WHERE
    ->  ( manager_id, department_id ) IN (
    ->          SELECT
    ->                  manager_id,
    ->                  department_id
    ->          FROM
    ->                  employees
    ->          WHERE
    ->                  employee_id IN ( 141, 174 )
    ->          )
    ->  AND
    ->  employee_id NOT IN ( 141, 174 );
+-------------+------------+---------------+
| employee_id | manager_id | department_id |
+-------------+------------+---------------+
|         142 |        124 |            50 |
|         143 |        124 |            50 |
|         144 |        124 |            50 |
|         196 |        124 |            50 |
|         197 |        124 |            50 |
|         198 |        124 |            50 |
|         199 |        124 |            50 |
|         175 |        149 |            80 |
|         176 |        149 |            80 |
|         177 |        149 |            80 |
|         179 |        149 |            80 |
+-------------+------------+---------------+
11 rows in set (0.00 sec)

2.3 HAVING 中的子查询

首先执行子查询 ，向主查询中的HAVING 子句返回结果

查询最低工资大于50号部门最低工资的部门id和其最低工资

sql 复制代码

mysql> SELECT
    ->  department_id,
    ->  MIN(salary)
    -> FROM
    ->  employees
    -> GROUP BY
    ->  department_id
    -> HAVING
    ->  MIN(salary) > (
    ->          SELECT MIN(salary)
    ->          FROM employees
    ->          WHERE department_id = 50
    ->  );
+---------------+-------------+
| department_id | MIN(salary) |
+---------------+-------------+
|          NULL |     7000.00 |
|            10 |     4400.00 |
|            20 |     6000.00 |
|            30 |     2500.00 |
|            40 |     6500.00 |
|            60 |     4200.00 |
|            70 |    10000.00 |
|            80 |     6100.00 |
|            90 |    17000.00 |
|           100 |     6900.00 |
|           110 |     8300.00 |
+---------------+-------------+
11 rows in set (0.00 sec)

2.4 CASE中的子查询

在CASE表达式中使用单列子查询

显式员工的employee_id,last_name和location。其中，若员工department_id与location_id为1800 的department_id相同，则location为'Canada'，其余则为'USA'

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  last_name,
    ->  (CASE
    ->          WHEN department_id = (
    ->                  SELECT department_id
    ->                  FROM departments
    ->                  WHERE location_id = 1800
    ->                  )
    ->                  THEN 'Canada'
    ->                  ELSE 'USA'
    ->  END) AS location
    -> FROM
    ->  employees;
+-------------+-------------+----------+
| employee_id | last_name   | location |
+-------------+-------------+----------+
|         100 | King        | USA      |
|         101 | Kochhar     | USA      |
|         ... | ...         | ...      |
|         201 | Hartstein   | Canada   |
|         202 | Fay         | Canada   |
|         203 | Mavris      | USA      |
|         204 | Baer        | USA      |
|         205 | Higgins     | USA      |
|         206 | Gietz       | USA      |
+-------------+-------------+----------+
(...为省略记录)
107 rows in set (0.00 sec)

2.5 子查询中的空值问题

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  job_id
    -> FROM
    ->  employees
    -> WHERE
    ->  job_id = (
    ->          SELECT job_id
    ->          FROM employees
    ->          WHERE last_name = 'Haas'
    ->  );
Empty set (0.00 sec)

2.5 非法使用子查询

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  last_name
    -> FROM
    ->  employees
    -> WHERE
    ->  salary = (
    ->          SELECT MIN( salary )
    ->          FROM employees
    ->          GROUP BY department_id
    ->  );
ERROR 1242 (21000): Subquery returns more than 1 row

子查询返回多行（多行子查询），使用单行比较操作符会报错

3. 多行子查询

也称为集合比较子查询
内查询返回多行
使用多行比较操作符

3.1 多行比较操作符

操作符	含义
IN	等于列表中的任意一个
ANY	需要和单行比较操作符一起使用，和子查询返回的某一个值比较
ALL	需要和单行比较操作符一起使用，和子查询返回的所有值比较
SOME	实际上是ANY的别名，作用相同，一般常使用ANY

3.2 代码示例

返回其它job_id中比job_id为'IT_PROG'部门任一工资低的员工的员工号、姓名、job_id 以及salary

sql 复制代码

mysql> SELECT
    ->  employee_id,
    ->  last_name,
    ->  job_id,
    ->  salary
    -> FROM
    ->  employees
    -> WHERE
    ->  salary < ANY(
    ->          SELECT salary
    ->          FROM employees
    ->          WHERE job_id = 'IT_PROG'
    ->  )
    ->  AND
    ->  job_id <> 'IT_PROG';
+-------------+-------------+------------+---------+
| employee_id | last_name   | job_id     | salary  |
+-------------+-------------+------------+---------+
|         110 | Chen        | FI_ACCOUNT | 8200.00 |
|         111 | Sciarra     | FI_ACCOUNT | 7700.00 |
|         112 | Urman       | FI_ACCOUNT | 7800.00 |
|         113 | Popp        | FI_ACCOUNT | 6900.00 |
|         ... | ...         | ...        |     ... |
+-------------+-------------+------------+---------+
(...为省略记录)
76 rows in set (0.00 sec)

查询平均工资最低的部门id

方法一：

sql 复制代码

mysql> SELECT
    ->  department_id
    -> FROM
    ->  employees
    -> GROUP BY
    ->  department_id
    -> HAVING
    ->  AVG(salary) <= ALL(
    ->          SELECT
    ->                  AVG(salary)
    ->          FROM
    ->                  employees
    ->          GROUP BY
    ->                  department_id
    ->  );
+---------------+
| department_id |
+---------------+
|            50 |
+---------------+
1 row in set (0.00 sec)

方法二：(重要思想：把子查询的查询结果当成一个新的表，写在 FROM 后面)

sql 复制代码

mysql> SELECT
    ->  department_id
    -> FROM
    ->  employees
    -> GROUP BY
    ->  department_id
    -> HAVING
    ->  AVG( salary ) = (
    ->          SELECT
    ->                  MIN( avg_sal )
    ->          FROM(
    ->                  SELECT AVG( salary ) avg_sal
    ->                  FROM employees
    ->                  GROUP BY department_id 
    ->          )dept_avg_sal
    ->  );
+---------------+
| department_id |
+---------------+
|            50 |
+---------------+
1 row in set (0.00 sec)

3.3 空值问题

sql 复制代码

mysql> SELECT
    ->  last_name
    -> FROM
    ->  employees
    -> WHERE
    ->  employee_id NOT IN (
    ->          SELECT manager_id
    ->          FROM employees
    ->  );
Empty set (0.00 sec)

子查询结果有NULL，可能不是想要的，就需要加过滤条件，去掉里面的NULL

4. 相关子查询

4.1 相关子查询执行流程

前面的案例均为不相关子查询

概念：如果子查询的执行依赖于外部查询，通常情况下都是因为子查询中的表用到了外部的表，并进行了条件关联，因此每执行一次外部查询，子查询都要重新计算一次，这样的子查询就称之为关联子查询

相关子查询按照一行接一行的顺序执行，主查询的每一行都执行一次子查询

说明：子查询中使用主查询中的列

4.2 代码示例

查询员工中工资大于本部门平均工资的员工的last_name,salary和其department_id

方法一：相关子查询

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  salary,
    ->  department_id
    -> FROM
    ->  employees e
    -> WHERE
    ->  salary > (
    ->          SELECT AVG(salary)
    ->          FROM employees
    ->          WHERE department_id = e.department_id
    ->  );
+-----------+----------+---------------+
| last_name | salary   | department_id |
+-----------+----------+---------------+
| King      | 24000.00 |            90 |
| Hunold    |  9000.00 |            60 |
| Ernst     |  6000.00 |            60 |
| Greenberg | 12000.00 |           100 |
| Faviet    |  9000.00 |           100 |
| Raphaely  | 11000.00 |            30 |
| Weiss     |  8000.00 |            50 |
| ...       |      ... |           ... |
+-----------+----------+---------------+
(...为省略记录)
38 rows in set (0.00 sec)

方法二：在 FROM 中使用子查询

sql 复制代码

mysql> SELECT
    ->  last_name,
    ->  salary,
    ->  e1.department_id
    -> FROM
    ->  employees e1,
    ->  ( SELECT
    ->      department_id,
    ->      AVG( salary ) dept_avg_sal
    ->    FROM employees
    ->    GROUP BY department_id
    ->  ) e2
    -> WHERE
    ->  e1.department_id = e2.department_id
    ->  AND e2.dept_avg_sal < e1.salary;
+-----------+----------+---------------+
| last_name | salary   | department_id |
+-----------+----------+---------------+
| Hartstein | 13000.00 |            20 |
| Raphaely  | 11000.00 |            30 |
| Weiss     |  8000.00 |            50 |
| Fripp     |  8200.00 |            50 |
| Kaufling  |  7900.00 |            50 |
| Vollman   |  6500.00 |            50 |
| ...       |      ... |           ... |
+-----------+----------+---------------+
(...为省略记录)
38 rows in set (0.00 sec)

from型的子查询：子查询是作为from的一部分，子查询要用()引起来，并且要给这个子查询取别名，把它当成一张"临时的虚拟的表"来使用

在ORDER BY 中使用子查询：

查询员工的id,salary,按照department_name 排序

sql 复制代码

SELECT
	employee_id,
	salary
FROM
	employees e
ORDER BY
	(
		SELECT department_name
		FROM departments d
		WHERE e.department_id = d.department_id
	) ASC;

结论：

SQL 复制代码

SELECT ...,...,...(存在聚合函数）
FROM ...(LEFT/RIGHT)JOIN...ON 多表的连接条件
WHERE 不包含聚合函数的过滤条件
GROUP BY ...,...
HAVING 包含聚合函数的过滤条件
ORDER BY ...,...(ASC/DESC)
LIMIT ...,...

除了GROUP BY 和 LIMIT 中不能写子查询，其他均可

若employees表中employee_id与job_history表中employee_id相同的数目不小于2，输出这些相同 id的员工的employee_id,last_name和其job_id

sql 复制代码

SELECT
	employee_id,
	last_name,
	job_id
FROM
	employees e
WHERE
	2 <= (
		SELECT COUNT(*)
		FROM job_history j
		WHERE e.employee_id = j.employee_id
	);

4.3 EXISTS 与 NOT EXISTS关键字

相关子查询中，有时会用，用来检查在子查询中是否存在满足条件的行

EXISTS
- 如果在子查询中不存在 满足条件的行：
  - 条件返回 FALSE
  - 继续在子查询中查找
- 如果在子查询中存在满足条件的行：
  - 不在子查询中继续查找
  - 条件返回 TRUE
NOT EXISTS 关键字表示如果不存在某种条件，则返回TRUE，否则返回FALSE （与EXISTS相反）

练习：

查询公司管理者的employee_id，last_name，job_id，department_id信息

方法一：自连接

sql 复制代码

SELECT
	DISTINCT 
	e1.employee_id,
	e1.last_name,
	e1.department_id
FROM
	employees e1 JOIN employees e2
WHERE
	e1.employee_id = e2.manager_id;

方法二：子查询

sql 复制代码

SELECT
	employee_id,
	last_name,
	department_id
FROM
	employees
WHERE
	employee_id IN (
		SELECT manager_id
		FROM employees
	);

方法三：使用EXISTS

sql 复制代码

SELECT
	employee_id,
	last_name,
	department_id
FROM
	employees e1
WHERE
	EXISTS(
		SELECT *
		FROM employees e2
		WHERE e1.employee_id = e2.manager_id
	);

解释：用 e1 的 employee_id 逐条与 e2 中的 manager_id 对比，找到一条就停止，返回 true ，未找到返回 false

查询departments表中，不存在于employees表中的部门的department_id和department_name

sql 复制代码

SELECT
	department_id,
	department_name 
FROM
	departments d 
WHERE
	NOT EXISTS ( 
		SELECT * 
		FROM employees 
		WHERE department_id = d.department_id 
	);

4.4 相关更新

sql 复制代码

UPDATE table1 alias1 
SET COLUMN = ( 
	SELECT expression 
	FROM table2 alias2 
	WHERE alias1.COLUMN = alias2.COLUMN 
	);

使用相关子查询依据一个表中的数据更新另一个表的数据

在employees中增加一个department_name字段，数据为员工对应的部门名称

sql 复制代码

# 1）
ALTER TABLE employees 
ADD (
		department_name VARCHAR2 ( 14 )
	);

# 2）
UPDATE employees e 
SET department_name = ( 
	SELECT department_name 
	FROM departments d 
	WHERE e.department_id = d.department_id 
	);

4.5 相关删除

sql 复制代码

DELETE 
FROM
	table1 alias1 
WHERE
	COLUMN operator ( 
		SELECT expression 
		FROM table2 alias2 
		WHERE alias1.COLUMN = alias2.COLUMN 
	);

使用相关子查询依据一个表中的数据删除另一个表的数据

删除表employees中，其与emp_history表皆有的数据

sql 复制代码

DELETE 
FROM
	employees e 
WHERE
	employee_id IN ( 
		SELECT employee_id 
		FROM emp_history 
		WHERE employee_id = e.employee_id 
	);

5. 自连接与子查询的效率问题

一般情况建议使用自连接，因为在许多 DBMS 的处理过程中，对于自连接的处理速度要比子查询快得多

理解：子查询实际上是通过未知表进行查询后的条件判断，而自连接是通过已知的自身数据表进行条件判断，因此在大部分 DBMS 中都对自连接处理进行了优化