MySQL优化之BTree索引使用规则
MySQL优化之BTree索引使用规则
从一道题开始分析：
假设某个表有一个联合索引（c1,c2,c3,c4）以下那个只能使用该联合索引的c1,c2,c3部分
explain select * from t where c1='a1' and c2='a2' and c4='a4' and c3='a3';
explain select * from t where c1='a1' and c2='a2' and c4='a4' order by c3;
explain select * from t where c1='a1' and c2='a2' and c4='a4' order by c1;
explain select * from t where c1='a1' and c4='a4'group by c3,c2;
explain select * from t where c1='a1' and c4='a4'group by c2,c3;
explain select * from t where c1='a1' and c4='a4'order by c2,c3;
explain select * from t where c1='a1' and c5='a5' order by c2,c3;
创建表后插入数据：
insert into t
values
('a1','a2','a3','a4','a5'),
('b1','b2','b3','b4','b5');
添加索引：
alter table t add index c1234(c1,c2,c3,c4);
mysql> explain select * from t where c1='a1' and c2='a2' and c4='a4' and c3='a3’;（MySQL执行时会改变其中的位置，c3会在c4的前边）
+----+-------------+-------+------+---------------+-------+---------+-------------------------+------+-----------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------------------------+------+-----------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 168 | const,const,const,const | 1 | Using index condition |
+----+-------------+-------+------+---------------+-------+---------+-------------------------+------+-----------------------+
1 行于数据集 (0.05 秒)
mysql> explain select * from t where c2='a1' and c3='a4’; (根据索引的左前缀规则，c2,c3部分没有索引）
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| 1 | SIMPLE | t | ALL | NULL | NULL | NULL | NULL | 4 | Using where |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
1 行于数据集 (0.08 秒)
mysql> explain select * from t where c1='a1' and c2='a2' and c4='a4' order by c3;（根据索引的左前缀规则，order by部分没有索引）
+----+-------------+-------+------+---------------+-------+---------+-------------+------+------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------------+------+------------------------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 84 | const,const | 1 | Using index condition; Using where |
+----+-------------+-------+------+---------------+-------+---------+-------------+------+------------------------------------+
1 行于数据集 (0.08 秒)
mysql> explain select * from t where c1='a1' and c2='a2' and c4='a4' order by c1;（根据索引的左前缀规则，order by使用了索引，可以存在where与order by同时使用索引的情况）
+----+-------------+-------+------+---------------+-------+---------+-------------+------+-----------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------------+------+-----------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 84 | const,const | 1 | Using index condition |
+----+-------------+-------+------+---------------+-------+---------+-------------+------+-----------------------+
1 行于数据集 (0.09 秒)
（以下四组对比，说明在group by 和order by 中，使用索引的顺序，可以避免使用临时表和filesort）
mysql> explain select * from t where c1='a1' and c4=‘a4' group by c3,c2;
+----+-------------+-------+------+---------------+-------+---------+-------+------+---------------------------------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------+------+---------------------------------------------------------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 42 | const | 1 | Using index condition; Using where; Using temporary; Using filesort |
+----+-------------+-------+------+---------------+-------+---------+-------+------+---------------------------------------------------------------------+
1 行于数据集 (0.09 秒)
mysql> explain select * from t where c1='a1' and c4='a4'group by c2,c3;
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 42 | const | 1 | Using index condition; Using where |
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
1 行于数据集 (0.12 秒)
mysql> explain select * from t where c1='a1' and c4='a4'order by c2,c3;
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 42 | const | 1 | Using index condition; Using where |
+----+-------------+-------+------+---------------+-------+---------+-------+------+------------------------------------+
1 行于数据集 (0.13 秒)
mysql> explain select * from t where c1='a1' and c5='a5' order by c3,c2;(使用的c1索引)
+----+-------------+-------+------+---------------+-------+---------+-------+------+----------------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+-------+---------+-------+------+----------------------------------------------------+
| 1 | SIMPLE | t | ref | c1234 | c1234 | 42 | const | 1 | Using index condition; Using where; Using filesort |
+----+-------------+-------+------+---------------+-------+---------+-------+------+----------------------------------------------------+
1 行于数据集 (0.15 秒)