CREATE TABLE `user_group` (  `id` int(11) NOT NULL auto_increment,  `uid` int(11) NOT NULL,  `group_id` int(11) NOT NULL,  PRIMARY KEY  (`id`),  KEY `uid` (`uid`),  KEY `group_id` (`group_id`),) ENGINE=InnoDB AUTO_INCREMENT=750366 DEFAULT CHARSET=utf8 
看AUTO_INCREMENT就知道數(shù)據(jù)并不多，75萬條。簡單的查詢：
SELECT SQL_NO_CACHE uid FROM user_group WHERE group_id = 245;   -- SQL_NO_CACHE 不使用緩存提示
Explain的結(jié)果是：+----+-------------+------------+------+---------------+----------+---------+-------+------+-------+| id | select_type | table      | type | possible_keys | key      | key_len | ref   | rows | Extra |+----+-------------+------------+------+---------------+----------+---------+-------+------+-------+|  1 | SIMPLE      | user_group | ref  | group_id      | group_id | 4       | const | 5544 |       |+----+-------------+------------+------+---------------+----------+---------+-------+------+-------+
看起來已經(jīng)用上索引了，數(shù)據(jù)分布上，group_id相同的比較多，uid散列的比較均勻，加索引的效果一般，試著加了一個多列索引：ALTER TABLE user_group ADD INDEX group_id_uid (group_id, uid);
這句SQL查詢的性能發(fā)生了巨大的提升，居然已經(jīng)可以跑到0.00s左右了。經(jīng)過優(yōu)化的SQL再結(jié)合真實的業(yè)務(wù)需求，也從之前2.2s下降到0.05s。再Explain一次+----+-------------+------------+------+-----------------------+--------------+---------+-------+------+-------------+| id | select_type | table      | type | possible_keys         | key          | key_len | ref   | rows | Extra       |+----+-------------+------------+------+-----------------------+--------------+---------+-------+------+-------------+|  1 | SIMPLE      | user_group | ref  | group_id,group_id_uid | group_id_uid | 4       | const | 5378 | Using index |+----+-------------+------------+------+-----------------------+--------------+---------+-------+------+-------------+
這種叫覆蓋索引（covering index），MySQL只需要通過索引就可以返回查詢所需要的數(shù)據(jù)，而不必在查到索引之后再去查詢數(shù)據(jù)，所以相當(dāng)快?。?/span>但是同時也要求所查詢的字段必須被索引所覆蓋到，在Explain的時候，輸出的Extra信息中如果有“Using Index”，就表示這條查詢使用了覆蓋索引。