mysql索引不生效的解决方法

存储引擎使用InnoDB。

一开始在主库查询，一直很好奇为什么索引不生效，切换到备库之后，发现备库是有效的。

开始考虑是不是因为索引出问题，后对索引重建，发现效率高了不少。

简单记录一下对比。

mysql> explain select * from runinfo where status in (0, 2, 1, 3, 4, 7, 9, 10);
+----+-------------+---------+-------+---------------+------+---------+------+----------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------+-------+---------------+------+---------+------+----------+-------------+
| 1 | SIMPLE | runinfo | All | status_2 | NULL | NULL | NULL | 2378055 | Using where |
+----+-------------+---------+-------+---------------+------+---------+------+----------+-------------+
row in set (0.00 sec)

上面是主库的执行计划。

对比一下备库的执行计划。

mysql> explain select * from runinfo where status in (0, 2, 1, 3, 4, 7, 9, 10);
+----+-------------+---------+-------+---------------+----------+---------+------+------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+---------+-------+---------------+----------+---------+------+------+-------------+
| 1 | SIMPLE | runinfo | range | status_2 | status_2 | 4 | NULL | 116 | Using where |
+----+-------------+---------+-------+---------------+----------+---------+------+------+-------------+
row in set (0.00 sec)

可以看出，备库在查询时适应到索引 status_2。

执行如下的命令之后，问题解决。

mysql> OPTIMIZE TABLE runinfo;
+------------------+----------+----------+-------------------------------------------------------------------+
| Table | Op | Msg_type | Msg_text |
+------------------+----------+----------+-------------------------------------------------------------------+
| schedule.runinfo | optimize | note | Table does not support optimize, doing recreate + analyze instead |
| schedule.runinfo | optimize | status | OK |
+------------------+----------+----------+-------------------------------------------------------------------+
rows in set (47.13 sec)

第二天来看，查询再一次变慢，有点好奇是不是有新数据写入导致索引不更新。