100W数据，测试复合索引

复合索引不是那么容易被catch到的。

两个查询条件都是等于的时候，才会被catch到。

mysql> select count(*) from tf_user_index where sex = 2 and score > 80;
+----------+
| count(*) |
+----------+
|  1261904 |
+----------+
1 row in set (10.65 sec)

mysql> select count(*) from tf_user where sex = 2 and score > 80;
+----------+
| count(*) |
+----------+
|  1261904 |
+----------+
1 row in set (4.38 sec)

mysql> explain select count(*) from tf_user_index where sex = 2 and score > 80;
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+
| id | select_type | table         | type | possible_keys | key  | key_len | ref   | rows    | Extra       |
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+
|  1 | SIMPLE      | tf_user_index | ref  | score,sex     | sex  | 1       | const | 4481227 | Using where |
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+
1 row in set (0.08 sec)

查询条件中，如果有大于号。那么优先抓取等于号对应的索引，也就是sex对应的索引。经过索引的一番折腾，查询时间反而更长了。

即便是把score放到前面，一样的效果。

mysql> explain select count(*) from tf_user_index where score> 80 and sex = 2;
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+
| id | select_type | table         | type | possible_keys | key  | key_len | ref   | rows    | Extra       |
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+
|  1 | SIMPLE      | tf_user_index | ref  | score,sex     | sex  | 1       | const | 4481227 | Using where |
+----+-------------+---------------+------+---------------+------+---------+-------+---------+-------------+

两个条件都为等于的时候，索引的效果就有点明显了。

mysql> select count(*) from tf_user_index where sex = 2 and score = 80;
+----------+
| count(*) |
+----------+
|    63230 |
+----------+
1 row in set (1.09 sec)

mysql> select count(*) from tf_user where sex = 2 and score = 80;
+----------+
| count(*) |
+----------+
|    63230 |
+----------+
1 row in set (2.61 sec)

mysql> explain select count(*) from tf_user_index where sex = 2 and score = 80;
+----+-------------+---------------+-------------+---------------+-----------+---------+------+--------+------------------------------------------------------+
| id | select_type | table         | type        | possible_keys | key       | key_len | ref  | rows   | Extra                                                |
+----+-------------+---------------+-------------+---------------+-----------+---------+------+--------+------------------------------------------------------+
|  1 | SIMPLE      | tf_user_index | index_merge | score,sex     | score,sex | 4,1     | NULL | 124004 | Using intersect(score,sex); Using where; Using index |
+----+-------------+---------------+-------------+---------------+-----------+---------+------+--------+------------------------------------------------------+
1 row in set (0.00 sec)

这个时候，并没有添加复合索引。

加了复合索引，如果查询条件是大于号，一样catch不到。

mysql> explain select count(*) from tf_user_index where sex = 2 and score > 80;
+----+-------------+---------------+------+---------------------+------+---------+-------+---------+-------------+
| id | select_type | table         | type | possible_keys       | key  | key_len | ref   | rows    | Extra       |
+----+-------------+---------------+------+---------------------+------+---------+-------+---------+-------------+
|  1 | SIMPLE      | tf_user_index | ref  | score,sex,score_sex | sex  | 1       | const | 4481227 | Using where |
+----+-------------+---------------+------+---------------------+------+---------+-------+---------+-------------+
1 row in set (0.01 sec)

mysql> select count(*) from tf_user_index where score > 80 and sex =2;
+----------+
| count(*) |
+----------+
|  1261904 |
+----------+
1 row in set (15.32 sec)

竟然执行了15秒之久。

这条sql语句可以优化一下，将sex也改为大于号。

mysql> select count(*) from tf_user_index where score > 80 and sex >1;
+----------+
| count(*) |
+----------+
|  1261904 |
+----------+
1 row in set (0.66 sec)

mysql> explain select count(*) from tf_user_index where score > 80 and sex >1;
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
| id | select_type | table         | type  | possible_keys       | key       | key_len | ref  | rows    | Extra                    |
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
|  1 | SIMPLE      | tf_user_index | range | score,sex,score_sex | score_sex | 4       | NULL | 4481227 | Using where; Using index |
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
1 row in set (0.00 sec)

这样就捕捉到了索引。

mysql> select count(*) from tf_user_index where score = 80 and sex = 1;
+----------+
| count(*) |
+----------+
|    62866 |
+----------+
1 row in set (0.01 sec)

mysql> explain select count(*) from tf_user_index where score = 80 and sex = 1;
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
| id | select_type | table         | type | possible_keys       | key       | key_len | ref         | rows   | Extra       |
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
|  1 | SIMPLE      | tf_user_index | ref  | score,sex,score_sex | score_sex | 5       | const,const | 124794 | Using index |
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
1 row in set (0.00 sec)

mysql> select count(*) from tf_user_index where  sex = 1 and score=80;
+----------+
| count(*) |
+----------+
|    62866 |
+----------+
1 row in set (0.01 sec)

mysql> explain select count(*) from tf_user_index where  sex = 1 and score=80;
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
| id | select_type | table         | type | possible_keys       | key       | key_len | ref         | rows   | Extra       |
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
|  1 | SIMPLE      | tf_user_index | ref  | score,sex,score_sex | score_sex | 5       | const,const | 124794 | Using index |
+----+-------------+---------------+------+---------------------+-----------+---------+-------------+--------+-------------+
1 row in set (0.00 sec)

顺序并不重要。

索引，复合索引，确实可以提供查询的速度。关键是，要能够捕捉到。要能够找寻它们捕捉的规律。理解它们执行的过程。

合理的分析查询的规律，合理的给表添加索引。分析常用的查询，分析常用的查询字段。通过explain字段来进行sql语句的分析，优化sql语句。

实践发现sex_score,score_sex查询的效果是一样的，关键是能否捕捉到。

mysql> explain select count(*) from tf_user_index where  sex = 2 and score>80;
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
| id | select_type | table         | type  | possible_keys       | key       | key_len | ref  | rows    | Extra                    |
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
|  1 | SIMPLE      | tf_user_index | range | score,sex,sex_score | sex_score | 5       | NULL | 2446346 | Using where; Using index |
+----+-------------+---------------+-------+---------------------+-----------+---------+------+---------+--------------------------+
1 row in set (0.00 sec)

mysql> select count(*) from tf_user_index where  sex = 2 and score>80;
+----------+
| count(*) |
+----------+
|  1261904 |
+----------+
1 row in set (0.31 sec)

我屮艸芔茻，sex_score竟然捕捉到了索引。看来顺序还是有所区别的。这个建索引还是多多的实验吧。孰能生巧。