MySQL之Join
参见MySQL(以5.1为例)中官方手册:MySQL官方手册-JOIN
假设有以下几个表
id | book |
---|---|
1 | java |
2 | c++ |
3 | php |
id | author |
---|---|
2 | zhang |
3 | wang |
4 | li |
author | year |
---|---|
zhang | 2003 |
ma | 2006 |
liu | 2011 |
Inner Join 内连接
将两个表中存在连接关系的字段,组成的记录集,叫做内连接。
内连接等价于
mysql> select t1.id as id,book,author from t1, t2 where t1.id=t2.id; +------+------+--------+ | id | book | author | +------+------+--------+ | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+ 2 rows in set (0.00 sec) mysql> select * from t1 inner join t2 using (id); +------+------+--------+ | id | book | author | +------+------+--------+ | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+ 2 rows in set (0.00 sec)
可以看出,两者是等价的。没有Using子句的Inner Join相当于是求两个表的笛卡尔积。
Cross Join 交叉连接
在Mysql中,Cross Join可以用逗号表达式表示,例如(t1, t2)。在Mysql中,Cross Join 和 Inner Join 是等价的,但是在标准SQL中,它们并不等价,Inner Join 用于带有on表达式的连接,反之用Cross Join。以下两个SQL语句是等价的。
Cross Join 指的是两个table的笛卡尔积。以下三句SQL是等价的。
mysql> select * from t1 inner join t2; mysql> select * from t1 cross join t2; mysql> select * from (t1, t2); mysql> select * from t1 nature join t2; 结果集: +------+------+------+--------+ | id | book | id | author | +------+------+------+--------+ | 1 | java | 2 | zhang | | 2 | c++ | 2 | zhang | | 3 | php | 2 | zhang | | 1 | java | 3 | wang | | 2 | c++ | 3 | wang | | 3 | php | 3 | wang | | 1 | java | 4 | li | | 2 | c++ | 4 | li | | 3 | php | 4 | li | +------+------+------+--------+
下面两句SQL也是等价的。
mysql> select * from table1 left join (table2, table3) on (table2.id = table1.id and table2.author = table3.author); mysql> select * from table1 left join (table2 cross join table3) on (table2.id = table1.id and table2.author = table3.author); 结果集: +------+------+------+--------+--------+------+ | id | book | id | author | author | year | +------+------+------+--------+--------+------+ | 1 | java | NULL | NULL | NULL | NULL | | 2 | c++ | 2 | zhang | zhang | 2003 | | 3 | php | NULL | NULL | NULL | NULL | +------+------+------+--------+--------+------+
Natural Join 自然连接
NATURAL [LEFT] JOIN:这个句子的作用相当于INNER JOIN,或者是在USING子句中包含了联结的表中所有公共字段的Left JOIN(左联结)。
也就是说:下面两个SQL是等价的。
mysql> select * from table1 natural join table2; mysql> select * from table1 inner join table2 using (id); 结果集: +------+------+--------+ | id | book | author | +------+------+--------+ | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+
同时,下面两个SQL也是等价的。
mysql> select * from table1 natural left join table2; mysql> select * from table1 left join table2 using(id); 结果集: +------+------+--------+ | id | book | author | +------+------+--------+ | 1 | java | NULL | | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+
Left Join 左外连接
左外连接A、B表的意思就是将表A中的全部记录和表B中字段连接形成的记录集,这里注意的是最后出来的记录集会包括表A的全部记录。
左连接表1,表二等价于右连接表二,表一。如下两个SQL是等价的:
mysql> select * from table1 left join table2 using (id); mysql> select * from table2 right join table1 using (id); 结果集: +------+------+--------+ | id | book | author | +------+------+--------+ | 1 | java | NULL | | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+
Right Join 右外连接
右外连接和左外连接是类似的。为了方便数据库便于访问,推荐使用左外连接代替右外连接。
Mysql表连接的一些注意事项:
1、两个表求差集的方法
如果求 左表 - 右表 的差集,使用类似下面的SQL:
SELECT left_tbl.* FROM left_tbl LEFT JOIN right_tbl ON left_tbl.id = right_tbl.id WHERE right_tbl.id IS NULL; 例如 mysql> select table1.* from table1 left join table2 using(id) where table2.id is null; +------+------+ | id | book | +------+------+ | 1 | java | +------+------+ 1 row in set (0.00 sec)
2、Using子句
Using子句可以使用On子句重写。但是使用Select * 查询出的结果有差别。以下两句话是等价的:
mysql> select id, book, author from table1 join table2 using (id);
mysql> select table1.id, book, author from table1 join table2 on table1.id=table2.id;
结果集:
+------+------+--------+
| id | book | author |
+------+------+--------+
| 2 | c++ | zhang |
| 3 | php | wang |
+------+------+--------+
但是下面两个有些许不同,使用on时候,重复的部分会被输出两次。
mysql> select * from table1 join table2 using (id); +------+------+--------+ | id | book | author | +------+------+--------+ | 2 | c++ | zhang | | 3 | php | wang | +------+------+--------+ 2 rows in set (0.00 sec) mysql> select * from table1 join table2 on table1.id=table2.id; +------+------+------+--------+ | id | book | id | author | +------+------+------+--------+ | 2 | c++ | 2 | zhang | | 3 | php | 3 | wang | +------+------+------+--------+ 2 rows in set (0.00 sec)
3、Straight Join的使用
STRAIGHT_JOIN 和 JOIN相似,除了大部分情况下,在使用STRAIGHT_JOIN时候,先读右表后读左表。而在大部分情况下是先读左表的。STRAIGHT_JOIN仅用于少数情况下的表连接性能优化,比如右表记录数目明显少于左表。
4、Mysql表连接的运算顺序
在MySQL 5.1版本中,INNER JOIN, CROSS JOIN, LEFT JOIN, 和RIGHT JOIN 比逗号表达式具有更高的优先级。
因此SQL1被解析成SQL3,而不是SQL2。
SQL1 : SELECT * FROM t1, t2 JOIN t3 ON (t1.i1 = t3.i3); SQL2 : SELECT * FROM (t1, t2) JOIN t3 ON (t1.i1 = t3.i3); SQL3 : SELECT * FROM t1, (t2 JOIN t3 ON (t1.i1 = t3.i3));
因此会报错,找不到i1列。因此以后在写这样的查询的时候,最好写明白,不要省略括号,这样能避免很多错误。
5、循环的自然连接
在MySQL 5.1版本中,SQL1等价于SQL3, 而在MySQL以前版本中,SQL1等价于SQL2。
SQL1 : SELECT ... FROM t1 NATURAL JOIN t2 NATURAL JOIN t3; SQL2 : SELECT ... FROM t1, t2, t3 WHERE t1.b = t2.b AND t2.c = t3.c; SQL3 : SELECT ... FROM t1, t2, t3 WHERE t1.b = t2.b AND t2.c = t3.c AND t1.a = t3.a;