深入浅出:MySQL的左连接、右连接、等值连接
三种连接的语法
为便于更多的技友快速读懂、理解,我们只讨论2张表对象进行连接操作的情况,大于2张表对象进行的连接操作原理也是一样的。
1.左连接(LEFT JOIN )
SELECT M.columnname……,N.* columnname…..
FROM left_table M LEFT JOIN right_table N ON M.columnname_join=N.columnname_join AND N.columnname=XXX
WHERE M.columnname=XXX…..
2.右连接(RIGHT JOIN)
SELECT M.columnname……,N.* columnname…..
FROM left_table M RIGHT JOIN right_table N ON M. columnname_join=N. columnname_join AND M. columnname=XXX
WHERE N.columnname=XXX…..
3.等值连接
SELECT M.columnname……,N.* columnname…..
FROM left_table M [INNER] JOIN right_table N ON M. columnname_join=N. columnname_join
WHERE M.columnname=XXX….. AND N.columnname=XXX….
或者
SELECT M.columnname……,N.* columnname…..
FROM left_table M , right_table N
WHERE M. columnname_join=N. columnname_join AND
M.columnname=XXX….. AND N.columnname=XXX….
说明:
A.左连接(LEFT JOIN )
ON字句连接条件,用于把2表中等值的记录连接在一起,但是不影响记录集的数量。若是表left_table中的某记录,无法在表right_table找到对应的记录,则此记录依然显示在记录集钟,只是表 right_table需要在查询显示的列的值用NULL替代;
ON字句连接条件中表right_table.columnname=XXX用于控制right_table表是否有符合要求的列值还是用NULL替换的方式显示在查询列中,不影响记录集的数量;
WHERE字句控制记录是否符合查询要求,不符合则过滤掉;
简单点说,就是 “left_table.columnname_join=right_table.columnname_join”用于统计两个表的等值连接的记录集,on子句
连接条件就是限制right_table的记录集大小,where子句的条件是限制最终结果集的大小。最终结果集的大小是显示“left_table“表的所有记录,和“left_table“匹配的”right_table“的记录显示”right_table“列值,不匹配的话,显示对应的”right_table“显示为null(最终结果集的大小是left_table刨除where条件的记录数)
B.右连接(RIGHT JOIN)
ON子句连接条件,用于把2表中等值的记录连接在一起,若是表right_table中的某记录,无法在表left_table找到对应的记录,则表 left_table需要在查询显示的列的值用NULL替代;
ON子句连接条件中表left_table.columnname=XXX用于控制left_table表是否有符合要求的列值,还是用NULL替换的方式显示在查询列表中;
WHERE字句控制记录是否符合查询要求,不符合则过滤掉;
简单点说,就是 “left_table.columnname_join=right_table.columnname_join”用于统计两个表的等值连接的记录集,on子句
连接条件就是限制left_table的记录集大小,where子句的条件是限制最终结果集的大小。最终结果集的大小是显示“right_table”表的所有记录,和“right_table“匹配的”left_table“的记录显示”left_table“列值,不匹配的话,显示对应的”left_table“显示为null(最终结果集的大小是left_table刨除where条件的记录数)
C. 等值连接
ON子句连接条件,不再与左连接或右连接的功效一样,除了作为2表记录匹配的条件外,还会起到过滤记录的作用,若left_table中记录无法在right_table中找到对应的记录,则会被过滤掉;
WHERE字句,不管是涉及表left_table、表right_table上的限制条件,还是涉及2表连接的条件,都会对记录集起到过滤作用,把不符合要求的记录刷选掉;
测试数据:
创建表
mysql> CREATE TABLE left_table(ID INT UNSIGNED NOT NULL AUTO_INCREMENT,
-> Username VARCHAR(40) NOT NULL,
-> Birthday DATETIME NOT NULL DEFAULT '0000-00-00 00:00:00',
-> CityID SMALLINT NOT NULL DEFAULT 0,
-> CreatDate TIMESTAMP NOT NULL DEFAULT '0000-00-00 00:00:00',
-> AlterDate TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
-> PRIMARY KEY(ID),
-> KEY idx_username(Username)
-> )ENGINE=InnoDB CHARACTER SET 'utf8' COLLATE 'utf8_general_ci';
Query OK, 0 rows affected (0.11 sec)
mysql>
mysql> CREATE TABLE right_table(UID INT UNSIGNED NOT NULL ,
-> CollectNum MEDIUMINT NOT NULL DEFAULT 0,
-> BuyNum MEDIUMINT NOT NULL DEFAULT 0,
-> SearchNum MEDIUMINT NOT NULL DEFAULT 0,
-> CreatDate TIMESTAMP NOT NULL DEFAULT '0000-00-00 00:00:00',
-> AlterDate TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
-> PRIMARY KEY(UID)
-> )ENGINE=InnoDB CHARACTER SET 'utf8' COLLATE 'utf8_general_ci';
Query OK, 0 rows affected (0.00 sec)
插入测试数据
执行如下sql十次
INSERT INTO left_table(Username,Birthday,CityID,CreatDate,AlterDate)
VALUES(CONCAT(SUBSTRING(RAND(),3,8),'@qq.com'),DATE_ADD(NOW(),INTERVAL -SUBSTRING(RAND(),3,2) YEAR),SUBSTRING(RAND(),3,2),DATE_ADD(NOW(),INTERVAL -SUBSTRING(RAND(),3,3) DAY),DATE_ADD(NOW(),INTERVAL -SUBSTRING(RAND(),3,2) DAY));
执行一次如下sql
INSERT INTO right_table
SELECT ID,SUBSTRING(RAND(),3,4) AS CollectNum,SUBSTRING(RAND(),3,2) AS BuyNum,SUBSTRING(RAND(),3,3) AS SearchNum,CreatDate,AlterDate FROM left_table WHERE ID%5=1;
查看基础数据:
mysql> select * from right_table;
+-----+------------+--------+-----------+---------------------+---------------------+
| UID | CollectNum | BuyNum | SearchNum | CreatDate | AlterDate |
+-----+------------+--------+-----------+---------------------+---------------------+
| 1 | 2545 | 78 | 171 | 2010-01-13 17:42:37 | 2011-12-16 17:42:37 |
| 6 | 4951 | 96 | 325 | 2009-11-28 17:42:55 | 2011-12-10 17:42:55 |
+-----+------------+--------+-----------+---------------------+---------------------+
2 rows in set (0.00 sec)
mysql> select * from left_table;
+----+-----------------+---------------------+--------+---------------------+---------------------+
| ID | Username | Birthday | CityID | CreatDate | AlterDate |
+----+-----------------+---------------------+--------+---------------------+---------------------+
| 1 | [email protected] | 1995-02-20 17:42:37 | 5 | 2010-01-13 17:42:37 | 2011-12-16 17:42:37 |
| 2 | [email protected] | 2004-02-20 17:42:49 | 39 | 2010-03-02 17:42:49 | 2012-01-10 17:42:49 |
| 3 | [email protected] | 1990-02-20 17:42:52 | 44 | 2010-08-28 17:42:52 | 2012-01-14 17:42:52 |
| 4 | [email protected] | 1989-02-20 17:42:53 | 32 | 2009-07-09 17:42:53 | 2011-12-03 17:42:53 |
| 5 | [email protected] | 2003-02-20 17:42:54 | 19 | 2010-03-25 17:42:54 | 2011-11-23 17:42:54 |
| 6 | [email protected] | 1987-02-20 17:42:55 | 11 | 2009-11-28 17:42:55 | 2011-12-10 17:42:55 |
| 7 | [email protected] | 1949-02-20 17:42:56 | 68 | 2010-09-17 17:42:56 | 2011-12-27 17:42:56 |
| 8 | [email protected] | 1967-02-20 17:42:57 | 59 | 2010-06-18 17:42:57 | 2012-01-24 17:42:57 |
| 9 | [email protected] | 1921-02-20 17:42:59 | 91 | 2009-11-23 17:42:59 | 2012-01-15 17:42:59 |
| 10 | [email protected] | 1940-02-20 17:43:00 | 51 | 2011-01-09 17:43:00 | 2012-01-02 17:43:00 |
+----+-----------------+---------------------+--------+---------------------+---------------------+
10 rows in set (0.00 sec)
mysql> select m.id,m.username from left_table m where m.id<=6;
+----+-----------------+
| id | username |
+----+-----------------+
| 1 | [email protected] |
| 2 | [email protected] |
| 3 | [email protected] |
| 4 | [email protected] |
| 5 | [email protected] |
| 6 | [email protected] |
+----+-----------------+
6 rows in set (0.02 sec)
左连接(left join)
编号:sql_1
mysql> select m.id,m.username,n.CollectNum,n.BuyNum from left_table m left join right_table n on m.id=n.uid where m.id <=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | 2545 | 78 |
| 2 | [email protected] | NULL | NULL |
| 3 | [email protected] | NULL | NULL |
| 4 | [email protected] | NULL | NULL |
| 5 | [email protected] | NULL | NULL |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
6 rows in set (0.02 sec)
编号:sql_2
mysql> select m.id,m.username,n.CollectNum,n.BuyNum from left_table m left join right_table n on m.id=n.uid and n.SearchNum>300 where m.id <=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | NULL | NULL |
| 2 | [email protected] | NULL | NULL |
| 3 | [email protected] | NULL | NULL |
| 4 | [email protected] | NULL | NULL |
| 5 | [email protected] | NULL | NULL |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
6 rows in set (0.00 sec)
sql_1和sql_2对比
1.两个结果集的记录数是一样的,left_table显示的列值是一样的
2.两个结果集中right_table相应列值是不一样的
为什么会不一样呢?
两个记录集的记录数是一样的,那是因为两个sql的where条件是一样,即最终的结果集的数量就会一样,right_table相应的列值不一样那是因为在on的条件中sql_2过滤掉了right_table的符合条件的记录。
右连接也就类似了,下面说下等值连接。
带join的等值连接
mysql> select m.id,m.username,n.CollectNum,n.BuyNum
-> from left_table m inner join right_table n on m.id=n.uid
-> where m.id<=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | 2545 | 78 |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
2 rows in set (0.00 sec)
mysql> select m.id,m.username,n.CollectNum,n.BuyNum from left_table m inner join right_table n on m.id=n.uid and n.SearchNum >300 where m.id<=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
1 row in set (0.01 sec)
从上面的可以看到,这里的on有了过滤最终结果集的大小的作用了
不带join的等值连接(这种方式我用着比较舒服,因为一直用oracle的)
mysql> select m.id,m.username,n.CollectNum,n.BuyNum
-> from left_table m,right_table n
-> where m.id=n.uid and m.id<=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | 2545 | 78 |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
2 rows in set (0.00 sec)
mysql>
常见错误用法的左连接
1.错把限制表right_table的条件,从ON子句中放到WHERE字句中,这样就会影响最终记录集大小
mysql> select m.id,m.username,n.CollectNum,n.BuyNum from left_table m left join right_table n on m.id=n.uid and n.SearchNum>300 where m.id <=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | NULL | NULL |
| 2 | [email protected] | NULL | NULL |
| 3 | [email protected] | NULL | NULL |
| 4 | [email protected] | NULL | NULL |
| 5 | [email protected] | NULL | NULL |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
6 rows in set (0.00 sec)
select m.id,m.username,n.CollectNum,n.BuyNum from left_table m ,right_table n where m.id=n.uid and m.id <=6 and n.SearchNum > 300;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
1 row in set (0.00 sec)
2.错把限制表left_table或称影响最终记录集的条件,从WHERE子句中放到ON字句中,这样就相当于增大的最终结果集记录数
select m.id,m.username,n.CollectNum,n.BuyNum from left_table m ,right_table n where m.id=n.uid and m.id <=6 and n.SearchNum > 300;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
1 row in set (0.00 sec)
mysql> select m.id,m.username,n.CollectNum,n.BuyNum from left_table m left join right_table n on m.id=n.uid and n.SearchNum>300 where m.id <=6;
+----+-----------------+------------+--------+
| id | username | CollectNum | BuyNum |
+----+-----------------+------------+--------+
| 1 | [email protected] | NULL | NULL |
| 2 | [email protected] | NULL | NULL |
| 3 | [email protected] | NULL | NULL |
| 4 | [email protected] | NULL | NULL |
| 5 | [email protected] | NULL | NULL |
| 6 | [email protected] | 4951 | 96 |
+----+-----------------+------------+--------+
6 rows in set (0.00 sec)