C++ 全栈知识体系

MySQL进阶 - 索引使用

索引的使用包含以下8个部分内容。

  1. 验证索引效率
  2. 最左前缀法则
  3. 范围查询
  4. 索引失效情况
  5. SQL提示
  6. 覆盖索引
  7. 前缀索引
  8. 单列索引与联合索引

验证索引效率

在讲解索引的使用原则之前，先通过一个简单的案例，来验证一下索引，看看是否能够通过索引来提升数据查询性能。

在演示的时候，使用之前准备的一张表 big_data , 在这张表中准备了1000w的记录。

这张表中id为主键，有主键索引，而其他字段是没有建立索引的。我们先来查询其中的一条记录，看里面的字段情况，执行如下SQL：

mysql> select * from big_data where id = 1;
+----+-------+------+---------------+
| id | name  | age  | email         |
+----+-------+------+---------------+
|  1 | test0 |   37 | test0@163.com |
+----+-------+------+---------------+
1 row in set (0.00 sec)
1
2
3
4
5
6
7

可以看到即使有1000w的数据,根据id进行数据查询,性能依然很快，因为主键id是有索引的。 那么接下来，我们再来根据 email 字段进行查询，执行如下SQL：

mysql> select * from big_data where email = 'test0@163.com';
+----+-------+------+---------------+
| id | name  | age  | email         |
+----+-------+------+---------------+
|  1 | test0 |   37 | test0@163.com |
+----+-------+------+---------------+
1 row in set (4.38 sec)
1
2
3
4
5
6
7

我们可以看到根据 email 字段进行查询，查询返回了一条数据，结果耗时 4.39sec，就是因为 email 没有索引，而造成查询效率很低。

那么我们可以针对于 email 字段，建立一个索引，建立了索引之后，我们再次根据 email 进行查询，再来看一下查询耗时情况。

创建索引：

mysql> create index idx_big_data_email on big_data(email);
Query OK, 0 rows affected (46.39 sec)
Records: 0  Duplicates: 0  Warnings: 0
1
2
3

然后再次执行相同的SQL语句，再次查看SQL的耗时。

ysql> select * from big_data where email = 'test0@163.com';
+----+-------+------+---------------+
| id | name  | age  | email         |
+----+-------+------+---------------+
|  1 | test0 |   37 | test0@163.com |
+----+-------+------+---------------+
1 row in set (0.02 sec)
1
2
3
4
5
6
7

email 字段建立了索引之后，查询性能大大提升。建立索引前后，查询耗时都不是一个数量级的。

最左前缀法则

如果索引关联了多列（联合索引），要遵守最左前缀法则，最左前缀法则指的是查询从索引的最左列开始，并且不跳过索引中的列。 如果跳跃某一列，索引将部分失效（后面的字段索引失效）。

以 tb_user 表为例，我们先来查看一下之前 tb_user 表所创建的索引。

在 tb_user 表中，有一个联合索引，这个联合索引涉及到三个字段，顺序分别为：profession，age，status。

对于最左前缀法则指的是，查询时，最左变的列必须存在，否则索引全部失效。而且中间不能跳过某一列，否则该列后面的字段索引将失效。

接下来，我们来演示几组案例，看一下具体的执行计划：

mysql> explain select * from tb_user where profession = '软件工程' and age = 31 and status= '0';
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys                      | key                   | key_len | ref               | rows | filtered | Extra |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat,idx_user_age | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------------+------+----------+-------+
1 row in set, 1 warning (0.02 sec)

mysql> explain select * from tb_user where profession = '软件工程' and age = 31;
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys                      | key                   | key_len | ref         | rows | filtered | Extra |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat,idx_user_age | idx_user_pro_age_stat | 38      | const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession = '软件工程';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | const |    4 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

而且由以上三组测试，我们也可以推测出profession字段索引长度为36、age字段索引长度为2、status字段索引长度为4。

mysql> explain select * from tb_user where age = 31 and status = '0';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   24 |     4.17 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where status = '0';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   24 |    10.00 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

而通过上面的这两组测试，我们也可以看到索引并未生效，原因是因为不满足最左前缀法则，联合索引最左边的列profession不存在。

mysql> explain select * from tb_user where profession = '软件工程' and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref   | rows | filtered | Extra                 |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | const |    4 |    10.00 | Using index condition |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

上述的SQL查询时，存在profession字段，最左边的列是存在的，索引满足最左前缀法则的基本条件。但是查询时，跳过了age这个列，所以后面的列索引是不会使用的，也就是索引部分生效，所以索 引的长度就是36。

mysql> explain select * from tb_user where age = 31 and status = '0' and profession = '软件工程';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

注意 ：最左前缀法则中指的最左边的列，是指在查询时，联合索引的最左边的字段(即是第一个字段)必须存在，与编写SQL时，条件编写的先后顺序无关。

范围查询

联合索引中，出现范围查询(>,<)，范围查询右侧的列索引失效。

mysql> explain select * from tb_user where profession = '软件工程' and age > 30 and status = '0';
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
| id | select_type | table   | partitions | type  | possible_keys         | key                   | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | range | idx_user_pro_age_stat | idx_user_pro_age_stat | 38      | NULL |    2 |    10.00 | Using index condition |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

当范围查询使用> 或 < 时，走联合索引了，但是索引的长度为38，就说明范围查询右边的status字段是没有走索引的。

mysql> explain select * from tb_user where profession = '软件工程' and age >= 30 and status = '0';
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
| id | select_type | table   | partitions | type  | possible_keys         | key                   | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | range | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | NULL |    2 |    10.00 | Using index condition |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

当范围查询使用>= 或 <= 时，走联合索引了，但是索引的长度为42，就说明所有的字段都是走索引的。

所以，在业务允许的情况下，尽可能的使用类似于 >= 或 <= 这类的范围查询，而避免使用 > 或 <。

索引失效情况

索引列运算

不要在索引列上进行运算操作， 索引将失效。

在tb_user表中，除了前面介绍的联合索引之外，还有一个索引，是phone字段的单列索引。

A. 当根据phone字段进行等值匹配查询时, 索引生效

mysql> explain select * from tb_user where phone = '17799990015';
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type  | possible_keys  | key            | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | const | idx_user_phone | idx_user_phone | 35      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

B. 当根据phone字段进行函数运算操作之后，索引失效。

mysql> explain select * from tb_user where substring(phone, 10, 2) = '15';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   24 |   100.00 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

字符串不加引号

字符串类型字段使用时，不加引号，索引将失效。

接下来，我们通过两组示例，来看看对于字符串类型的字段，加单引号与不加单引号的区别：

第一组：

mysql> explain select * from tb_user where profession = '软件工程' and age = 31 and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession = '软件工程' and age = 31 and status = 0;
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-----------------------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref         | rows | filtered | Extra                 |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 38      | const,const |    1 |    10.00 | Using index condition |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------+------+----------+-----------------------+
1 row in set, 2 warnings (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

第二组：

mysql> explain select * from tb_user where phone = '17799990015';
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type  | possible_keys  | key            | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | const | idx_user_phone | idx_user_phone | 35      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+-------+----------------+----------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where phone = 17799990015;
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys  | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | idx_user_phone | NULL | NULL    | NULL |   24 |    10.00 | Using where |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
1 row in set, 3 warnings (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

经过上面两组示例，我们会明显的发现，如果字符串不加单引号，对于查询结果，没什么影响，但是数据库存在隐式类型转换，索引将失效。

模糊查询

如果仅仅是尾部模糊匹配，索引不会失效。如果是头部模糊匹配，索引失效。

接下来，我们来看一下这三条SQL语句的执行效果，查看一下其执行计划：

由于下面查询语句中，都是根据profession字段查询，符合最左前缀法则，联合索引是可以生效的，我们主要看一下，模糊查询时，%加在关键字之前，和加在关键字之后的影响。

mysql> explain select * from tb_user where profession like '软件%';
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
| id | select_type | table   | partitions | type  | possible_keys         | key                   | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | range | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | NULL |    4 |   100.00 | Using index condition |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
1 row in set, 1 warning (0.01 sec)

mysql> explain select * from tb_user where profession like '%工程';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   24 |    11.11 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession like '%工%';
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | NULL          | NULL | NULL    | NULL |   24 |    11.11 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

经过上述的测试，我们发现，在like模糊查询中，在关键字后面加%，索引可以生效。而如果在关键字前面加了%，索引将会失效。

or连接条件

用or分割开的条件， 如果or前的条件中的列有索引，而后面的列中没有索引，那么涉及的索引都不会被用到。

由于age没有索引，所以即使id、phone有索引，索引也会失效。所以需要针对于age也要建立索引。

mysql> explain select * from tb_user where id = 10 or age = 23;
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | PRIMARY       | NULL | NULL    | NULL |   24 |    13.75 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where phone = '17799990017' or age = 23;
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys  | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | idx_user_phone | NULL | NULL    | NULL |   24 |    19.00 | Using where |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

然后，我们可以对age字段建立索引。建立了索引之后，我们再次执行上述的SQL语句，看看前后执行计划的变化。

mysql> create index idx_user_age on tb_user(age);
Query OK, 0 rows affected (0.04 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> explain select * from tb_user where id = 10 or age = 23;
+----+-------------+---------+------------+-------------+----------------------+----------------------+---------+------+------+----------+------------------------------------------------+
| id | select_type | table   | partitions | type        | possible_keys        | key                  | key_len | ref  | rows | filtered | Extra                                          |
+----+-------------+---------+------------+-------------+----------------------+----------------------+---------+------+------+----------+------------------------------------------------+
|  1 | SIMPLE      | tb_user | NULL       | index_merge | PRIMARY,idx_user_age | PRIMARY,idx_user_age | 4,2     | NULL |    3 |   100.00 | Using union(PRIMARY,idx_user_age); Using where |
+----+-------------+---------+------------+-------------+----------------------+----------------------+---------+------+------+----------+------------------------------------------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where phone = '17799990017' or age = 23;
+----+-------------+---------+------------+-------------+-----------------------------+-----------------------------+---------+------+------+----------+-------------------------------------------------------+
| id | select_type | table   | partitions | type        | possible_keys               | key                         | key_len | ref  | rows | filtered | Extra                                                 |
+----+-------------+---------+------------+-------------+-----------------------------+-----------------------------+---------+------+------+----------+-------------------------------------------------------+
|  1 | SIMPLE      | tb_user | NULL       | index_merge | idx_user_phone,idx_user_age | idx_user_phone,idx_user_age | 35,2    | NULL |    3 |   100.00 | Using union(idx_user_phone,idx_user_age); Using where |
+----+-------------+---------+------------+-------------+-----------------------------+-----------------------------+---------+------+------+----------+-------------------------------------------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

最终，我们发现，当or连接的条件，左右两侧字段都有索引时，索引才会生效。

数据分布影响

如果MySQL评估使用索引比全表更慢，则不使用索引。

mysql> explain select * from tb_user where phone >= '17799990005';
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys  | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | idx_user_phone | NULL | NULL    | NULL |   24 |    79.17 | Using where |
+----+-------------+---------+------------+------+----------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where phone >= '17799990020';
+----+-------------+---------+------------+-------+----------------+----------------+---------+------+------+----------+-----------------------+
| id | select_type | table   | partitions | type  | possible_keys  | key            | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+---------+------------+-------+----------------+----------------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | range | idx_user_phone | idx_user_phone | 35      | NULL |    4 |   100.00 | Using index condition |
+----+-------------+---------+------------+-------+----------------+----------------+---------+------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

经过测试我们发现，相同的SQL语句，只是传入的字段值不同，最终的执行计划也完全不一样，这是为什么呢？

就是因为MySQL在查询时，会评估使用索引的效率与走全表扫描的效率，如果走全表扫描更快，则放弃索引，走全表扫描。因为索引是用来索引少量数据的，如果通过索引查询返回大批量的数据，则还不如走全表扫描来的快，此时索引就会失效。

接下来，我们再来看看 is null 与 is not null 操作是否走索引。

mysql> explain select * from tb_user where profession is null;
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref   | rows | filtered | Extra                 |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | const |    1 |   100.00 | Using index condition |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession is not null;
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys         | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | idx_user_pro_age_stat | NULL | NULL    | NULL |   24 |   100.00 | Using where |
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
1
2
3
4
5
6
7
8
9
10
11
12
13
14

然后，再次执行上述的两条SQL，查看SQL语句的执行计划。

mysql> explain select * from tb_user where profession is null;
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys         | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ALL  | idx_user_pro_age_stat | NULL | NULL    | NULL |   24 |   100.00 | Using where |
+----+-------------+---------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession is not null;
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
| id | select_type | table   | partitions | type  | possible_keys         | key                   | key_len | ref  | rows | filtered | Extra                 |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
|  1 | SIMPLE      | tb_user | NULL       | range | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | NULL |    1 |   100.00 | Using index condition |
+----+-------------+---------+------------+-------+-----------------------+-----------------------+---------+------+------+----------+-----------------------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

最终我们看到，一模一样的SQL语句，先后执行了两次，结果查询计划是不一样的，为什么会出现这种现象，这是和数据库的数据分布有关系。查询时MySQL会评估，走索引快，还是全表扫描快，如果全表 扫描更快，则放弃索引走全表扫描。 因此，is null 、is not null是否走索引，得具体情况具体分析，并不是固定的。

总结

  1. 在索引列上进行运算操作，索引将失效。如：explain select * from tb_user where substring(phone, 10, 2) = '15';
  2. 字符串类型字段使用时，不加引号，索引将失效。如：explain select * from tb_user where phone = 17799990015;，此处phone的值没有加引号
  3. 模糊查询中，如果仅仅是尾部模糊匹配，索引不会是失效；如果是头部模糊匹配，索引失效。如：explain select * from tb_user where profession like '%工程';，前后都有 % 也会失效。
  4. 用 or 分割开的条件，如果 or 其中一个条件的列没有索引，那么涉及的索引都不会被用到。
  5. 如果 MySQL 评估使用索引比全表更慢，则不使用索引。

SQL 提示

是优化数据库的一个重要手段，简单来说，就是在SQL语句中加入一些人为的提示来达到优化操作的目的。

A. 执行SQL : explain select * from tb_user where profession = '软件工程';

B. 执行SQL，创建profession的单列索引：create index idx_user_pro on tb_user(profession);

C. 创建单列索引后，再次执行A中的SQL语句，查看执行计划，看看到底走哪个索引。

mysql> explain select * from tb_user where profession = '软件工程';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> create index idx_user_pro on tb_user(profession);
Query OK, 0 rows affected (0.03 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> explain select * from tb_user where profession = '软件工程';
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys                      | key                   | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat,idx_user_pro | idx_user_pro_age_stat | 36      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+------------------------------------+-----------------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.01 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

测试结果，我们可以看到，possible_keys中 idx_user_pro_age_sta,idx_user_pro 这两个索引都可能用到，最终MySQL选择了idx_user_pro_age_sta索引。这是MySQL自动选择的结果。

那么，我们能不能在查询的时候，自己来指定使用哪个索引呢？ 答案是肯定的，此时就可以借助于MySQL的SQL提示来完成。 接下来，介绍一下SQL提示。

SQL提示，是优化数据库的一个重要手段，简单来说，就是在SQL语句中加入一些人为的提示来达到优化操作的目的。

  1. use index: 建议MySQL使用哪一个索引完成此次查询（仅仅是建议，mysql内部还会再次进行评估）。
  2. ignore index: 忽略指定的索引。
  3. force index: 强制使用索引。

例如，使用索引： explain select * from tb_user use index(idx_user_pro) where profession="软件工程"; 不使用哪个索引： explain select * from tb_user ignore index(idx_user_pro) where profession="软件工程"; 必须使用哪个索引： explain select * from tb_user force index(idx_user_pro) where profession="软件工程";

测试结果：

mysql> explain select * from tb_user use index(idx_user_pro) where profession="软件工程";
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys | key          | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro  | idx_user_pro | 36      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user ignore index(idx_user_pro) where profession="软件工程";
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 36      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user force index(idx_user_pro) where profession="软件工程";
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys | key          | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro  | idx_user_pro | 36      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+---------------+--------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

覆盖索引&回表查询

尽量使用覆盖索引（查询使用了索引，并且需要返回的列，在该索引中已经全部能找到），减少 select *。

我们来看一组SQL的执行计划，看看执行计划的差别，然后再来具体做一个解析。

explain select id, profession from tb_user where profession = '软件工程' and age =31 and status = '0';
explain select id,profession,age,status from tb_user where profession = '软件工程' and age = 31 and status = '0';
explain select id,profession,age,status,name from tb_user where profession = '软件工程' and age = 31 and status = '0';
explain select * from tb_user where profession = '软件工程' and age = 31 and status = '0';
1
2
3
4

执行结果：

mysql> explain select id, profession from tb_user where profession = '软件工程' and age =31 and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra       |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | Using index |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select id,profession,age,status from tb_user where profession = '软件工程' and age = 31 and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra       |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | Using index |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select id,profession,age,status,name from tb_user where profession = '软件工程' and age = 31 and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from tb_user where profession = '软件工程' and age = 31 and status = '0';
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
| id | select_type | table   | partitions | type | possible_keys         | key                   | key_len | ref               | rows | filtered | Extra |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | ref  | idx_user_pro_age_stat | idx_user_pro_age_stat | 42      | const,const,const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+------+-----------------------+-----------------------+---------+-------------------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

explain 中 extra 字段含义： using index condition：查找使用了索引，但是需要回表查询数据 using where; using index;：查找使用了索引，但是需要的数据都在索引列中能找到，所以不需要回表查询

  • 如果在聚集索引中直接能找到对应的行，则直接返回行数据，只需要一次查询，哪怕是select *；
  • 如果在辅助索引中找聚集索引，如select id, name from xxx where name='xxx';，也只需要通过辅助索引(name)查找到对应的id，返回name和name索引对应的id即可，只需要一次查询；
  • 如果是通过辅助索引查找其他字段，则需要回表查询，如select id, name, gender from xxx where name='xxx';

所以尽量不要用select *，容易出现回表查询，降低效率，除非有联合索引包含了所有字段

面试题：一张表，有四个字段（id, username, password, status），由于数据量大，需要对以下SQL语句进行优化，该如何进行才是最优方案： select id, username, password from tb_user where username='itcast';

解：给username和password字段建立联合索引，则不需要回表查询，直接覆盖索引

前缀索引

当字段类型为字符串（varchar, text等）时，有时候需要索引很长的字符串，这会让索引变得很大，查询时，浪费大量的磁盘IO，影响查询效率，此时可以只降字符串的一部分前缀，建立索引，这样可以大大节约索引空间，从而提高索引效率。

语法： create index idx_xxxx on table_name(columnn(n));

前缀长度：可以根据索引的选择性来决定，而选择性是指不重复的索引值（基数）和数据表的记录总数的比值，索引选择性越高则查询效率越高，唯一索引的选择性是1，这是最好的索引选择性，性能也是最好的。

求选择性公式：

select count(distinct email) / count(*) from tb_user;
select count(distinct substring(email, 1, 5)) / count(*) from tb_user;
1
2

执行结果：

mysql> select count(distinct email) / count(*) from tb_user;
+----------------------------------+
| count(distinct email) / count(*) |
+----------------------------------+
|                           1.0000 |
+----------------------------------+
1 row in set (0.00 sec)

mysql> select count(distinct substring(email, 1, 5)) / count(*) from tb_user;
+---------------------------------------------------+
| count(distinct substring(email, 1, 5)) / count(*) |
+---------------------------------------------------+
|                                            0.9583 |
+---------------------------------------------------+
1 row in set (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

为tb_user表的email字段，建立长度为5的前缀索引。

mysql> create index idx_email_5 on tb_user(email(5));
Query OK, 0 rows affected (0.04 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> select * from tb_user where email = 'lvbu666@163.com';
+----+--------+-------------+-----------------+------------+------+--------+--------+---------------------+
| id | name   | phone       | email           | profession | age  | gender | status | createtime          |
+----+--------+-------------+-----------------+------------+------+--------+--------+---------------------+
|  1 | 吕布   | 17799990000 | lvbu666@163.com | NULL       |   23 | 1      | 6      | 2001-02-02 00:00:00 |
+----+--------+-------------+-----------------+------------+------+--------+--------+---------------------+
1 row in set (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11

前缀索引的查询流程:

单列索引&联合索引

单列索引：即一个索引只包含单个列

联合索引：即一个索引包含了多个列

在业务场景中，如果存在多个查询条件，考虑针对于查询字段建立索引时，建议建立联合索引，而非单列索引。

单列索引情况：

mysql> explain select id, phone, name from tb_user where phone = '17799990010' and name = '韩信';
+----+-------------+---------+------------+-------+------------------------------+----------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type  | possible_keys                | key            | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+-------+------------------------------+----------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | const | idx_user_phone,idx_user_name | idx_user_phone | 35      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+-------+------------------------------+----------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7

通过上述执行计划我们可以看出来，在and连接的两个字段 phone、name上都是有单列索引的，但是最终mysql只会选择一个索引phone，也就是说，只能走一个字段的索引，此时是会回表查询的。

紧接着，我们再来创建一个phone和name字段的联合索引来查询一下执行计划。

mysql> create unique index idx_user_phone_name on tb_user(phone,name);
Query OK, 0 rows affected (0.04 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> explain select id, phone, name from tb_user where phone = '17799990010' and name = '韩信';
+----+-------------+---------+------------+-------+--------------------------------------------------+----------------+---------+-------+------+----------+-------+
| id | select_type | table   | partitions | type  | possible_keys                                    | key            | key_len | ref   | rows | filtered | Extra |
+----+-------------+---------+------------+-------+--------------------------------------------------+----------------+---------+-------+------+----------+-------+
|  1 | SIMPLE      | tb_user | NULL       | const | idx_user_phone,idx_user_phone_name,idx_user_name | idx_user_phone | 35      | const |    1 |   100.00 | NULL  |
+----+-------------+---------+------------+-------+--------------------------------------------------+----------------+---------+-------+------+----------+-------+
1 row in set, 1 warning (0.00 sec)
1
2
3
4
5
6
7
8
9
10
11

此时，查询时，就走了联合索引，而在联合索引中包含 phone、name的信息，在叶子节点下挂的是对应的主键id，所以查询是无需回表查询的。

在业务场景中，如果存在多个查询条件，考虑针对于查询字段建立索引时，建议建立联合索引，而非单列索引。

如果查询使用的是联合索引，具体的结构示意图如下：

索引设计原则

  1. 针对于数据量较大，且查询比较频繁的表建立索引
  2. 针对于常作为查询条件（where）、排序（order by）、分组（group by）操作的字段建立索引
  3. 尽量选择区分度高的列作为索引，尽量建立唯一索引，区分度越高，使用索引的效率越高
  4. 如果是字符串类型的字段，字段长度较长，可以针对于字段的特点，建立前缀索引
  5. 尽量使用联合索引，减少单列索引，查询时，联合索引很多时候可以覆盖索引，节省存储空间，避免回表，提高查询效率
  6. 要控制索引的数量，索引并不是多多益善，索引越多，维护索引结构的代价就越大，会影响增删改的效率
  7. 如果索引列不能存储NULL值，请在创建表时使用NOT NULL约束它。当优化器知道每列是否包含NULL值时，它可以更好地确定哪个索引最有效地用于查询

← MySQL进阶 - SQL性能分析 →