在MariaDB数据库中，UNION运算符用于组合两个或更多SELECT语句的结果集。它删除各种SELECT语句之间的重复行。

语法

SELECT expression1, expression2, ... expression_n  
FROM tables  
[WHERE conditions]  
UNION [DISTINCT]  
SELECT expression1, expression2, ... expression_n  
FROM tables  
[WHERE conditions];

注意：UNION运算符中的每个SELECT语句在具有相似数据类型的结果集中必须具有相同数量的字段。

1. 使用UNION操作符返回单个字段

下面来看一个从多个SELECT语句返回一个字段的UNION运算符的例子。(两个表具有相同的公共字段)。

假设有两张表:students和teachers。对应的表结构和数据如下 –

students表中的数据：

MariaDB [testdb]> select * from students;
+------------+--------------+-----------------+----------------+
| student_id | student_name | student_address | admission_date |
+------------+--------------+-----------------+----------------+
|          1 | Maxsu        | Haikou          | 2017-01-07     |
|          2 | JMaster      | Beijing         | 2016-05-07     |
|          3 | Mahesh       | Guangzhou       | 2016-06-07     |
|          4 | Kobe         | Shanghai        | 2016-02-07     |
|          5 | Blaba        | Shenzhen        | 2016-08-07     |
+------------+--------------+-----------------+----------------+
5 rows in set (0.00 sec)

teachers表中的数据：

USE testdb;
CREATE TABLE teachers(  
    teacher_id INT NOT NULL AUTO_INCREMENT,  
    name VARCHAR(100) NOT NULL,  
    address VARCHAR(40) NOT NULL,  
    admission_date DATE,  
    PRIMARY KEY ( teacher_id )
);
-- 插入数据
INSERT INTO teachers  
(teacher_id, name, address, admission_date)  
VALUES(1,'陈小祥','上海','2013-06-07 00:00:00');

INSERT INTO teachers  
(teacher_id, name, address, admission_date)  
VALUES(2,'张天经','广州','2013-08-08 00:00:00');

INSERT INTO teachers  
(teacher_id, name, address, admission_date)  
VALUES(3,'李四光','三亚','2014-09-07 00:00:00');

经过上创建和插入数据，现在teachers表中拥有以下数据记录 –

MariaDB [testdb]> select * from teachers;
+------------+--------+---------+----------------+
| teacher_id | name   | address | admission_date |
+------------+--------+---------+----------------+
|          1 | 陈小祥 | 上海    | 2013-06-07     |
|          2 | 张天经 | 广州    | 2013-08-08     |
|          3 | 李四光 | 三亚    | 2014-09-07     |
+------------+--------+---------+----------------+
3 rows in set (0.08 sec)

假设现在要查询所有学生和老师的姓名，以及他们的地址，可参考以下查询语句 –

SELECT student_name AS name, student_address as address
FROM Students
UNION  
SELECT name,address
FROM teachers;

执行上面查询语句，得到以下结果 –

MariaDB [testdb]> SELECT student_name AS name, student_address as address
    -> FROM Students
    -> UNION
    -> SELECT name,address
    -> FROM teachers;
+---------+-----------+
| name    | address   |
+---------+-----------+
| Maxsu   | Haikou    |
| JMaster | Beijing   |
| Mahesh  | Guangzhou |
| Kobe    | Shanghai  |
| Blaba   | Shenzhen  |
| 陈小祥  | 上海      |
| 张天经  | 广州      |
| 李四光  | 三亚      |
+---------+-----------+
8 rows in set (0.05 sec)

它显示了两个表中并删除重复的name列的值。

2. 具有ORDER BY子句的UNION运算符

使用ORDER BY子句的UNION运算符从两个表中检索多个列。参考以下语句 –

SELECT student_id, student_name  
FROM Students  
WHERE student_name = 'Maxsu'  
UNION  
SELECT teacher_id, address
FROM teachers  
WHERE teacher_id > 1 
ORDER BY 1;

执行上面示例代码，得到以下结果 –

MariaDB [testdb]> SELECT student_id, student_name
    -> FROM Students
    -> WHERE student_name = 'Maxsu'
    -> UNION
    -> SELECT teacher_id, address
    -> FROM teachers
    -> WHERE teacher_id > 1
    -> ORDER BY 1;
+------------+--------------+
| student_id | student_name |
+------------+--------------+
|          1 | Maxsu        |
|          2 | 广州         |
|          3 | 三亚         |
+------------+--------------+
3 rows in set (0.00 sec)

3. 一些其它的用法

具有不同列名的表之间的UNION：

(SELECT e_name AS name, email FROM employees)
UNION
(SELECT c_name AS name, email FROM customers);

指定UNION的全局顺序并限制总行数：

(SELECT name, email FROM employees)
UNION
(SELECT name, email FROM customers)
ORDER BY name LIMIT 10;

添加一个常数行：

(SELECT 'John Doe' AS name, 'john.doe@example.net' AS email)
UNION
(SELECT name, email FROM customers);

不同的类型：

SELECT CAST('x' AS CHAR(1)) UNION SELECT REPEAT('y',4);
+----------------------+
| CAST('x' AS CHAR(1)) |
+----------------------+
| x                    |
| yyyy                 |
+----------------------+

按照每个SELECT使用排序列的顺序返回结果：

(SELECT 1 AS sort_column, e_name AS name, email FROM employees)
UNION
(SELECT 2, c_name AS name, email FROM customers) ORDER BY sort_column;