mysql loose index scan的实现

  Loose index scan

各种资料关于loose index scan的解释很拗口，其实等同于oracle的index skip scan。

对于复合索引(sex, id)，前导列sex的值只有两个：M和F，

(`F',98)

(`F',100)

(`F',102)

(`F',104)

(`M',101)

(`M',103)

(`M',105)

则索引可以逻辑的分成两个子索引

当执行select  * from t where id = N时，等价于

select  * from t where id = N and sex =‘M’

union

select  * from t where id = N and sex = 'F'

进而有效利用到索引进行查找，如果换成mysql，则只能执行全表扫描；
不考虑字段选择性的前提下，即便将sql写为select  * from t where id = N and sex in(‘M’,'F')，pre-5.6在索引扫描时也不会使用到id列过滤，而是在读取出行记录后再进行id=N的筛选。

但是5.0之后group by在特定条件下可能使用到loose index scan，以下是来自http://explainextended.com/2010/05/08/max-and-min-on-a-composite-index/

的案例。

CREATE TABLE log_table (

id INT NOT NULL PRIMARY KEY,

log_machine VARCHAR(20) NOT NULL,

log_time DATETIME NOT NULL

) ENGINE=InnoDB DEFAULT CHARSET=utf8;

CREATE INDEX ix_log_machine_time ON log_table (log_machine, log_time);

1

SELECT  MAX(log_time) FROM log_table；

SELECT MAX(log_time) FROM log_table WHERE log_machine IN ('Machine 1')；

这两条sql都只需一次index seek便可返回，源于索引的有序排序，优化器意识到min/max位于最左/右块，从而避免范围扫描；

extra显示Select tables optimized away ；

2

SELECT MAX(log_time) FROM log_table WHERE log_machine IN (‘Machine 1’,’Machine 2’,’Machine 3’,’Machine 4’);

执行计划type 为range(extra显示using where; using index)，即执行索引范围扫描，先读取所有满足log_machine约束的记录，然后对其遍历找出max value；

改进

SELECT MAX(log_time) FROM log_table WHERE log_machine IN (‘Machine 1’,’Machine 2’,’Machine 3’,’Machine 4’)  group by log_machine order by 1 desc limit 1;

这满足group by选择loose index scan的要求，执行计划的extra显示using index for group-by，执行效果等值于

SELECT MAX(log_time) FROM log_table WHERE log_machine IN (‘Machine 1’)

Union

SELECT MAX(log_time) FROM log_table WHERE log_machine IN (‘Machine 2’)

…..

即对每个log_machine执行loose index scan，rows从原来的82636下降为16(该表总共1,000,000条记录)。

Group by何时使用loose index scan?

适用条件：

1  针对单表操作

2  Group by使用索引的最左前缀列

3  只支持聚集函数min()/max()

4  Where条件出现的列必须为=constant操作 ， 没出现在group by中的索引列必须使用constant

5  不支持前缀索引，即部分列索引 ，如index(c1(10))

执行计划的extra应该显示using index for group-by

假定表t1有个索引idx(c1,c2,c3)

SELECT c1, c2 FROM t1 GROUP BY c1, c2;

SELECT DISTINCT c1, c2 FROM t1;

SELECT c1, MIN(c2) FROM t1 GROUP BY c1;

SELECT c1, c2 FROM t1 WHERE c1 < const GROUP BY c1, c2;

SELECT MAX(c3), MIN(c3), c1, c2 FROM t1 WHERE c2 > const GROUP BY c1, c2;

SELECT c2 FROM t1 WHERE c1 < const GROUP BY c1, c2;

SELECT c1, c2 FROM t1 WHERE c3 = const GROUP BY c1, c2

SELECT c1, c3 FROM t1 GROUP BY c1, c2;--无法使用松散索引

而SELECT c1, c3 FROM t1  where c3= const GROUP BY c1, c2;则可以

紧凑索引扫描tight index scan

Group by在无法使用loose index scan，还可以选择tight，若两者都不可选，则只能借助临时表；

扫描索引时，须读取所有满足条件的索引键，要么是全索引扫描，要么是范围索引扫描；

Group by的索引列不连续；或者不是从最左前缀开始，但是where条件里出现最左列；

SELECT c1, c2, c3 FROM t1 WHERE c2 = 'a' GROUP BY c1, c3;

SELECT c1, c2, c3 FROM t1 WHERE c1 = 'a' GROUP BY c2, c3;

http://blog.itpub.net/15480802/viewspace-757562/ 

5.6的改进

事实上，5.6的index condition push down可以弥补loose index scan缺失带来的性能损失。

KEY(age,zip)

mysql> explain SELECT  name FROM people WHERE age BETWEEN 18 AND 20 AND zip IN (12345,12346, 12347);

+----+-------------+--------+-------+---------------+------+---------+------+-------+-------------+

| id | select_type | table  | type  | possible_keys | key  | key_len | ref  | rows  | Extra       |

+----+-------------+--------+-------+---------------+------+---------+------+-------+-------------+

|  1 | SIMPLE      | people | range | age           | age  | 4       | NULL | 90556 | Using where |

+----+-------------+--------+-------+---------------+------+---------+------+-------+-------------+

1 row in set (0.01 sec)

根据key_len=4可以推测出sql只用到索引的第一列，即先通过索引查出满足age (18,20)的行记录，然后从server层筛选出满足zip约束的行；

pre-5.6，对于复合索引，只有当引导列使用"="时才有机会在索引扫描时使用到后面的索引列。

mysql> explain SELECT  name FROM people WHERE age=18 AND zip IN (12345,12346, 12347);

+----+-------------+--------+-------+---------------+------+---------+------+------+-------------+

| id | select_type | table  | type  | possible_keys | key  | key_len | ref  | rows | Extra       |

+----+-------------+--------+-------+---------------+------+---------+------+------+-------------+

|  1 | SIMPLE      | people | range | age           | age  | 8       | NULL |    3 | Using where |

+----+-------------+--------+-------+---------------+------+---------+------+------+-------------+

1 row in set (0.00 sec)

对比一下查询效率

mysql> SELECT  sql_no_cache name FROM people WHERE age=19 AND zip IN (12345,12346, 12347);

+----------------------------------+

| name                             |

+----------------------------------+

| 888ba838661aff00bbbce114a2a22423 |

+----------------------------------+

1 row in set (0.06 sec)

mysql> SELECT  SQL_NO_CACHE name FROM people WHERE age BETWEEN 18 AND 22 AND zip IN (12345,12346, 12347);

+----------------------------------+

| name                             |

+----------------------------------+

| ed4481336eb9adca222fd404fa15658e |

| 888ba838661aff00bbbce114a2a22423 |

+----------------------------------+

2 rows in set (1 min 56.09 sec)

对于第二条sql，可以使用union改写，

mysql> SELECT  name FROM people WHERE age=18 AND zip IN (12345,12346, 12347)

    -> UNION ALL

    -> SELECT  name FROM people WHERE age=19 AND zip IN (12345,12346, 12347)

    -> UNION ALL

    -> SELECT  name FROM people WHERE age=20 AND zip IN (12345,12346, 12347)

    -> UNION ALL

    -> SELECT  name FROM people WHERE age=21 AND zip IN (12345,12346, 12347)

    -> UNION ALL

-> SELECT  name FROM people WHERE age=22 AND zip IN (12345,12346, 12347);

而mysql5.6引入了index condition pushdown，从优化器层面解决了此类问题。

http://www.percona.com/blog/2006/08/10/using-union-to-implement-loose-index-scan-to-mysql/

http://blog.itpub.net/15480802/viewspace-1117157/

 

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/15480802/viewspace-1427989/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/15480802/viewspace-1427989/