MySQL的JOIN（一）：用法

時間 2019-11-07

標籤 mysql join 用法欄目 MySQL 简体版

原文原文鏈接

JOIN的含義就如英文單詞「join」同樣，鏈接兩張表，大體分爲內鏈接，外鏈接，右鏈接，左鏈接，天然鏈接。這裏描述先甩出一張用爛了的圖，而後插入測試數據。測試

測試數據表:優化

CREATE TABLE t_blog(
    id INT PRIMARY KEY AUTO_INCREMENT,
    title VARCHAR(50),
    typeId INT
);

INSERT INTO t_blog VALUES(NULL,'aaa',1),
   (NULL,'bbb',2),(NULL,'ccc',3),(NULL,'ddd',4),(NULL,'eee',4),
   (NULL,'fff',3),(NULL,'ggg',2),(NULL,'hhh',NULL),(NULL,'iii',NULL),(NULL,'jjj',NULL);

CREATE TABLE t_type(
   id INT PRIMARY KEY AUTO_INCREMENT,
   NAME VARCHAR(20)
);

INSERT INTO t_type VALUES (NULL,'C++'),
  (NULL,'C'),(NULL,'JAVA'),(NULL,'C#'),(NULL,'JVASCRIPT');

SELECT * FROM t_blog;
    +----+-------+--------+
    | id | title | typeId |
    +----+-------+--------+
    |  1 | aaa   |      1 |
    |  2 | bbb   |      2 |
    |  3 | ccc   |      3 |
    |  4 | ddd   |      4 |
    |  5 | eee   |      4 |
    |  6 | fff   |      3 |
    |  7 | ggg   |      2 |
    |  8 | hhh   |   NULL |
    |  9 | iii   |   NULL |
    | 10 | jjj   |   NULL |
    +----+-------+--------+

SELECT * FROM t_type;
    +----+------------+
    | id | name       |
    +----+------------+
    |  1 | C++        |
    |  2 | C          |
    |  3 | Java       |
    |  4 | C#         |
    |  5 | Javascript |
    +----+------------+

笛卡爾積：CROSS JOIN

要理解各類JOIN首先要理解笛卡爾積。笛卡爾積就是將A表的每一條記錄與B表的每一條記錄強行拼在一塊兒。因此，若是A表有n條記錄，B表有m條記錄，笛卡爾積產生的結果就會產生n*m條記錄。下面的例子，t_blog有10條記錄，t_type有5條記錄，全部他們倆的笛卡爾積有50條記錄。有五種產生笛卡爾積的方式以下。spa

SELECT * FROM t_blog CROSS JOIN t_type;
SELECT * FROM t_blog INNER JOIN t_type;
SELECT * FROM t_blog,t_type;
SELECT * FROM t_blog NATURE JOIN t_type;

+----+-------+--------+----+------------+
    | id | title | typeId | id | name       |
    +----+-------+--------+----+------------+
    |  1 | aaa   |      1 |  1 | C++        |
    |  1 | aaa   |      1 |  2 | C          |
    |  1 | aaa   |      1 |  3 | Java       |
    |  1 | aaa   |      1 |  4 | C#         |
    |  1 | aaa   |      1 |  5 | Javascript |
    |  2 | bbb   |      2 |  1 | C++        |
    |  2 | bbb   |      2 |  2 | C          |
    |  2 | bbb   |      2 |  3 | Java       |
    |  2 | bbb   |      2 |  4 | C#         |
    |  2 | bbb   |      2 |  5 | Javascript |
    |  3 | ccc   |      3 |  1 | C++        |
    |  3 | ccc   |      3 |  2 | C          |
    |  3 | ccc   |      3 |  3 | Java       |
    |  3 | ccc   |      3 |  4 | C#         |
    |  3 | ccc   |      3 |  5 | Javascript |
    |  4 | ddd   |      4 |  1 | C++        |
    |  4 | ddd   |      4 |  2 | C          |
    |  4 | ddd   |      4 |  3 | Java       |
    |  4 | ddd   |      4 |  4 | C#         |
    |  4 | ddd   |      4 |  5 | Javascript |
    |  5 | eee   |      4 |  1 | C++        |
    |  5 | eee   |      4 |  2 | C          |
    |  5 | eee   |      4 |  3 | Java       |
    |  5 | eee   |      4 |  4 | C#         |
    |  5 | eee   |      4 |  5 | Javascript |
    |  6 | fff   |      3 |  1 | C++        |
    |  6 | fff   |      3 |  2 | C          |
    |  6 | fff   |      3 |  3 | Java       |
    |  6 | fff   |      3 |  4 | C#         |
    |  6 | fff   |      3 |  5 | Javascript |
    |  7 | ggg   |      2 |  1 | C++        |
    |  7 | ggg   |      2 |  2 | C          |
    |  7 | ggg   |      2 |  3 | Java       |
    |  7 | ggg   |      2 |  4 | C#         |
    |  7 | ggg   |      2 |  5 | Javascript |
    |  8 | hhh   |   NULL |  1 | C++        |
    |  8 | hhh   |   NULL |  2 | C          |
    |  8 | hhh   |   NULL |  3 | Java       |
    |  8 | hhh   |   NULL |  4 | C#         |
    |  8 | hhh   |   NULL |  5 | Javascript |
    |  9 | iii   |   NULL |  1 | C++        |
    |  9 | iii   |   NULL |  2 | C          |
    |  9 | iii   |   NULL |  3 | Java       |
    |  9 | iii   |   NULL |  4 | C#         |
    |  9 | iii   |   NULL |  5 | Javascript |
    | 10 | jjj   |   NULL |  1 | C++        |
    | 10 | jjj   |   NULL |  2 | C          |
    | 10 | jjj   |   NULL |  3 | Java       |
    | 10 | jjj   |   NULL |  4 | C#         |
    | 10 | jjj   |   NULL |  5 | Javascript |
    +----+-------+--------+----+------------+

內鏈接：INNER JOIN

內鏈接INNER JOIN是最經常使用的鏈接操做。從數學的角度講就是求兩個表的交集，從笛卡爾積的角度講就是從笛卡爾積中挑出ON子句條件成立的記錄。有INNER JOIN，WHERE（等值鏈接），STRAIGHT_JOIN,JOIN(省略INNER)四種寫法。至於哪一種好我會在MySQL的JOIN（二）：優化講述。示例以下。設計

SELECT * FROM t_blog INNER JOIN t_type ON t_blog.typeId=t_type.id;
SELECT * FROM t_blog,t_type WHERE t_blog.typeId=t_type.id;
SELECT * FROM t_blog STRAIGHT_JOIN t_type ON t_blog.typeId=t_type.id; --注意STRIGHT_JOIN有個下劃線
SELECT * FROM t_blog JOIN t_type ON t_blog.typeId=t_type.id;

    +----+-------+--------+----+------+
    | id | title | typeId | id | name |
    +----+-------+--------+----+------+
    |  1 | aaa   |      1 |  1 | C++  |
    |  2 | bbb   |      2 |  2 | C    |
    |  7 | ggg   |      2 |  2 | C    |
    |  3 | ccc   |      3 |  3 | Java |
    |  6 | fff   |      3 |  3 | Java |
    |  4 | ddd   |      4 |  4 | C#   |
    |  5 | eee   |      4 |  4 | C#   |
    +----+-------+--------+----+------+

左鏈接：LEFT JOIN

左鏈接LEFT JOIN的含義就是求兩個表的交集外加左表剩下的數據。依舊從笛卡爾積的角度講，就是先從笛卡爾積中挑出ON子句條件成立的記錄，而後加上左表中剩餘的記錄（見最後三條）。code

SELECT * FROM t_blog LEFT JOIN t_type ON t_blog.typeId=t_type.id;

    +----+-------+--------+------+------+
    | id | title | typeId | id   | name |
    +----+-------+--------+------+------+
    |  1 | aaa   |      1 |    1 | C++  |
    |  2 | bbb   |      2 |    2 | C    |
    |  7 | ggg   |      2 |    2 | C    |
    |  3 | ccc   |      3 |    3 | Java |
    |  6 | fff   |      3 |    3 | Java |
    |  4 | ddd   |      4 |    4 | C#   |
    |  5 | eee   |      4 |    4 | C#   |
    |  8 | hhh   |   NULL | NULL | NULL |
    |  9 | iii   |   NULL | NULL | NULL |
    | 10 | jjj   |   NULL | NULL | NULL |
    +----+-------+--------+------+------+

右鏈接：RIGHT JOIN

同理右鏈接RIGHT JOIN就是求兩個表的交集外加右表剩下的數據。再次從笛卡爾積的角度描述，右鏈接就是從笛卡爾積中挑出ON子句條件成立的記錄，而後加上右表中剩餘的記錄（見最後一條）。blog

SELECT * FROM t_blog RIGHT JOIN t_type ON t_blog.typeId=t_type.id;

    +------+-------+--------+----+------------+
    | id   | title | typeId | id | name       |
    +------+-------+--------+----+------------+
    |    1 | aaa   |      1 |  1 | C++        |
    |    2 | bbb   |      2 |  2 | C          |
    |    3 | ccc   |      3 |  3 | Java       |
    |    4 | ddd   |      4 |  4 | C#         |
    |    5 | eee   |      4 |  4 | C#         |
    |    6 | fff   |      3 |  3 | Java       |
    |    7 | ggg   |      2 |  2 | C          |
    | NULL | NULL  |   NULL |  5 | Javascript |
    +------+-------+--------+----+------------+

外鏈接：OUTER JOIN

外鏈接就是求兩個集合的並集。從笛卡爾積的角度講就是從笛卡爾積中挑出ON子句條件成立的記錄，而後加上左表中剩餘的記錄，最後加上右表中剩餘的記錄。另外MySQL不支持OUTER JOIN，可是咱們能夠對左鏈接和右鏈接的結果作UNION操做來實現。ip

SELECT * FROM t_blog LEFT JOIN t_type ON t_blog.typeId=t_type.id
    UNION
    SELECT * FROM t_blog RIGHT JOIN t_type ON t_blog.typeId=t_type.id;

    +------+-------+--------+------+------------+
    | id   | title | typeId | id   | name       |
    +------+-------+--------+------+------------+
    |    1 | aaa   |      1 |    1 | C++        |
    |    2 | bbb   |      2 |    2 | C          |
    |    7 | ggg   |      2 |    2 | C          |
    |    3 | ccc   |      3 |    3 | Java       |
    |    6 | fff   |      3 |    3 | Java       |
    |    4 | ddd   |      4 |    4 | C#         |
    |    5 | eee   |      4 |    4 | C#         |
    |    8 | hhh   |   NULL | NULL | NULL       |
    |    9 | iii   |   NULL | NULL | NULL       |
    |   10 | jjj   |   NULL | NULL | NULL       |
    | NULL | NULL  |   NULL |    5 | Javascript |
    +------+-------+--------+------+------------+

USING子句

MySQL中鏈接SQL語句中，ON子句的語法格式爲：table1.column_name = table2.column_name。當模式設計對聯接表的列採用了相同的命名樣式時，就可使用 USING 語法來簡化 ON 語法，格式爲：USING(column_name)。
因此，USING的功能至關於ON，區別在於USING指定一個屬性名用於鏈接兩個表，而ON指定一個條件。另外，SELECT *時，USING會去除USING指定的列，而ON不會。實例以下。博客

SELECT * FROM t_blog INNER JOIN t_type ON t_blog.typeId =t_type.id;
    +----+-------+--------+----+------+
    | id | title | typeId | id | name |
    +----+-------+--------+----+------+
    |  1 | aaa   |      1 |  1 | C++  |
    |  2 | bbb   |      2 |  2 | C    |
    |  7 | ggg   |      2 |  2 | C    |
    |  3 | ccc   |      3 |  3 | Java |
    |  6 | fff   |      3 |  3 | Java |
    |  4 | ddd   |      4 |  4 | C#   |
    |  5 | eee   |      4 |  4 | C#   |
    +----+-------+--------+----+------+


    SELECT * FROM t_blog INNER JOIN t_type USING(typeId);
    ERROR 1054 (42S22): Unknown column 'typeId' in 'from clause'

    -- 這裏t_blog的typeId與t_type的id不一樣名，沒法用Using，這裏用id代替下。
    SELECT * FROM t_blog INNER JOIN t_type ON t_blog.id = t_type.id;
    +----+-------+--------+----+-----------+
    | id | title | typeId | id | name      |
    +----+-------+--------+----+-----------+
    |  1 | aaa   |      1 |  1 | C++       |
    |  2 | bbb   |      2 |  2 | C         |
    |  3 | ccc   |      3 |  3 | JAVA      |
    |  4 | ddd   |      4 |  4 | C#        |
    |  5 | eee   |      4 |  5 | JVASCRIPT |
    +----+-------+--------+----+-----------+

    SELECT * FROM t_blog INNER JOIN t_type USING(id); 
    +----+-------+--------+------------+
    | id | title | typeId | name       |
    +----+-------+--------+------------+
    |  1 | aaa   |      1 | C++        |
    |  2 | bbb   |      2 | C          |
    |  3 | ccc   |      3 | Java       |
    |  4 | ddd   |      4 | C#         |
    |  5 | eee   |      4 | Javascript |
    +----+-------+--------+------------+

天然鏈接：NATURE JOIN

天然鏈接就是USING子句的簡化版，它找出兩個表中相同的列做爲鏈接條件進行鏈接。有左天然鏈接，右天然鏈接和普通天然鏈接之分。在t_blog和t_type示例中，兩個表相同的列是id，因此會拿id做爲鏈接條件。
另外千萬分清下面三條語句的區別。
天然鏈接:SELECT * FROM t_blog NATURAL JOIN t_type; //NATURAL 爲關鍵字,表示天然鏈接
笛卡爾積:SELECT * FROM t_blog NATURA JOIN t_type; //NATURA 爲表的別名
笛卡爾積:SELECT * FROM t_blog NATURE JOIN t_type;//NATURE 爲表的別名數學

SELECT * FROM t_blog NATURAL JOIN t_type;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog,t_type WHERE t_blog.id=t_type.id;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog INNER JOIN t_type ON t_blog.id=t_type.id;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog INNER JOIN t_type USING(id);

    +----+-------+--------+------------+
    | id | title | typeId | name       |
    +----+-------+--------+------------+
    |  1 | aaa   |      1 | C++        |
    |  2 | bbb   |      2 | C          |
    |  3 | ccc   |      3 | Java       |
    |  4 | ddd   |      4 | C#         |
    |  5 | eee   |      4 | Javascript |
    +----+-------+--------+------------+

    SELECT * FROM t_blog NATURAL LEFT JOIN t_type;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog LEFT JOIN t_type ON t_blog.id=t_type.id;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog LEFT JOIN t_type USING(id);

    +----+-------+--------+------------+
    | id | title | typeId | name       |
    +----+-------+--------+------------+
    |  1 | aaa   |      1 | C++        |
    |  2 | bbb   |      2 | C          |
    |  3 | ccc   |      3 | Java       |
    |  4 | ddd   |      4 | C#         |
    |  5 | eee   |      4 | Javascript |
    |  6 | fff   |      3 | NULL       |
    |  7 | ggg   |      2 | NULL       |
    |  8 | hhh   |   NULL | NULL       |
    |  9 | iii   |   NULL | NULL       |
    | 10 | jjj   |   NULL | NULL       |
    +----+-------+--------+------------+

    SELECT * FROM t_blog NATURAL RIGHT JOIN t_type;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog RIGHT JOIN t_type ON t_blog.id=t_type.id;
    SELECT t_blog.id,title,typeId,t_type.name FROM t_blog RIGHT JOIN t_type USING(id);

    +----+------------+-------+--------+
    | id | name       | title | typeId |
    +----+------------+-------+--------+
    |  1 | C++        | aaa   |      1 |
    |  2 | C          | bbb   |      2 |
    |  3 | Java       | ccc   |      3 |
    |  4 | C#         | ddd   |      4 |
    |  5 | Javascript | eee   |      4 |
    +----+------------+-------+--------+

補充

博客開頭給出的第一張圖除去講了的內鏈接、左鏈接、右鏈接、外鏈接，還有一些特殊的韋恩圖，這裏補充一下。it

SELECT * FROM t_blog LEFT JOIN t_type ON t_blog.typeId=t_type.id
    WHERE t_type.id IS NULL;
    +----+-------+--------+------+------+
    | id | title | typeId | id   | name |
    +----+-------+--------+------+------+
    |  8 | hhh   |   NULL | NULL | NULL |
    |  9 | iii   |   NULL | NULL | NULL |
    | 10 | jjj   |   NULL | NULL | NULL |
    +----+-------+--------+------+------+
    SELECT * FROM t_blog RIGHT JOIN t_type ON t_blog.typeId=t_type.id
    WHERE t_blog.id IS NULL;
    +------+-------+--------+----+------------+
    | id   | title | typeId | id | name       |
    +------+-------+--------+----+------------+
    | NULL | NULL  |   NULL |  5 | Javascript |
    +------+-------+--------+----+------------+
    SELECT * FROM t_blog LEFT JOIN t_type ON t_blog.typeId=t_type.id
    WHERE t_type.id IS NULL
    UNION
    SELECT * FROM t_blog RIGHT JOIN t_type ON t_blog.typeId=t_type.id
    WHERE t_blog.id IS NULL;
    +------+-------+--------+------+------------+
    | id   | title | typeId | id   | name       |
    +------+-------+--------+------+------------+
    |    8 | hhh   |   NULL | NULL | NULL       |
    |    9 | iii   |   NULL | NULL | NULL       |
    |   10 | jjj   |   NULL | NULL | NULL       |
    | NULL | NULL  |   NULL |    5 | Javascript |
    +------+-------+--------+------+------------+

這篇博客發現有點「孔乙己：茴字的四種寫法的感受」，但仍是有收穫的。