1)myisam表:mysql> explain select * from a where id=1 or uid =2;
+----+-------------+-------+-------------+---------------+-------------+---------+------+------+---------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+-------------+---------------+-------------+---------+------+------+---------------------------------------+
| 1 | SIMPLE | a | index_merge | PRIMARY,uid | PRIMARY,uid | 4,4 | NULL | 2 |Using union(PRIMARY,uid);Using where |
+----+-------------+-------+-------------+---------------+-------------+---------+------+------+---------------------------------------+
1 row in set (0.00 sec)
在mysql5.0和更新的版本中,会同时使用这两个单列索引进行扫描,然后合并结果。索引合并需要耗费大量cpu和内存资源在算法的缓存,和并和排序上。
2)innodb表:(alter table a engine=innodb)
mysql> explain select * from a where id=1 or uid =2;
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| 1 | SIMPLE | a | ALL | PRIMARY,uid | NULL | NULL | NULL | 5 | Using where |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
1 row in set (0.00 sec)
2 .必须所有的or条件都必须是独立索引:
+-------+----------------------------------------------------------------------------------------------------------------------
| Table | Create Table
+-------+----------------------------------------------------------------------------------------------------------------------
| a | CREATE TABLE `a` (
`id` int(1) NOT NULL AUTO_INCREMENT,
`uid` int(11) NOT NULL,
`aNum` char(20) DEFAULT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=6 DEFAULT CHARSET=latin1 |
+-------+----------------------------------------------------------------------------------------------------------------------
1 row in set (0.00 sec)
explain查看:
mysql> explain select * from a where id=1 or uid =2;
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
| 1 | SIMPLE | a | ALL | PRIMARY | NULL | NULL | NULL | 5 | Using where |
+----+-------------+-------+------+---------------+------+---------+------+------+-------------+
1 row in set (0.00 sec)
全表扫描了。
3. 用UNION替换OR (适用于索引列)
通常情况下, 用UNION替换WHERE子句中的OR将会起到较好的效果. 对索引列使用OR将造成全表扫描.
注意, 以上规则只针对多个索引列有效. 如果有column没有被索引, 查询效率可能会因为你没有选择OR而降低.
在下面的例子中, LOC_ID 和REGION上都建有索引.
高效:
select loc_id , loc_desc , region from location where loc_id = 10 union select loc_id , loc_desc , regionfrom location where region = "melbourne"and loc_id != 10
ログイン後にコピー
低效:
select loc_id , loc desc , region from location where loc_id = 10 or region = "melbourne"
ログイン後にコピー
如果你坚持要用OR, 那就需要返回记录最少的索引列写在最前面.
通过查询表中值得分布情况(看看where条件的各个分支对应的数据基数有多大),确定那个值放在前边。select sum(loc_id = 10),sum(region="melbourne") from location/G************************1.row***********************SUM(loc_id = 10): 7992SUM(region="melbourne"): 30
ログイン後にコピー
通过检测发现应该把region="melbourne"放在前边。
4. 用in来替换or
这是一条简单易记的规则,但是实际的执行效果还须检验,在oracle8i下,两者的执行路径似乎是相同的.
低效:
select…. from location where loc_id = 10 or loc_id = 20 or loc_id = 30
高效
select… from location where loc_in in (10,20,30);