Mysql的一条SQL优化(一)

来源:互联网 发布:2015年度十大网络剧 编辑:程序博客网 时间:2024/04/30 09:36
开发反应做压测时并发上不去,有条SQL执行非常慢,于是explain下:
MariaDB [db_vip]> 
MariaDB [db_vip]> explain select            ext_id, mid, msource_id, msource_type, referee, mobile, tel, mpassword,status         from m_db_vip     where  (msource_id='xx1391434680574433' and msource_type=1 )     or ( mobile='1391434680574433'   and msource_type=1);
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
| id   | select_type | table     | type | possible_keys | key  | key_len | ref  | rows  | Extra       |
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
|    1 | SIMPLE      | m_db_vip | ALL  | NULL          | NULL | NULL    | NULL | 86987 | Using where |
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
1 row in set (0.00 sec)
MariaDB [db_vip]> show status like '%cost%';
+-----------------+--------------+
| Variable_name   | Value        |
+-----------------+--------------+
| Last_query_cost | 18006.399000 |
+-----------------+--------------+
1 row in set (0.00 sec)
从执行计划中可以看出,这个SQL对表这个表m_db_vip进行了全表扫描,而表m_db_vip共有约10万行记录,Mysql优化器评估出的成本为18006,
where条件的三列中msource_id与mobile都是选择性很强的列,msource_type的选择性弱,


首先对列msource_id创建一个索引:
MariaDB [db_vip]> create index i_m_db_vip_1 on m_db_vip(msource_id); 
Query OK, 0 rows affected (1.18 sec)
Records: 0  Duplicates: 0  Warnings: 0


再次执行SQL,发现Mysql并没有用上msource_id的索引(还是Oracle的CBO智能呀!)
MariaDB [db_vip]> explain select            ext_id, mid, msource_id, msource_type, referee, mobile, tel, mpassword,status         from m_db_vip     where  (msource_id='xx1391434680574433' and msource_type=1 )     or ( mobile='1391434680574433'   and msource_type=1);
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
| id   | select_type | table     | type | possible_keys | key  | key_len | ref  | rows  | Extra       |
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
|    1 | SIMPLE      | m_db_vip | ALL  | i_m_db_vip_1 | NULL | NULL    | NULL | 86987 | Using where |
+------+-------------+-----------+------+---------------+------+---------+------+-------+-------------+
1 row in set (0.01 sec)


MariaDB [db_vip]> show status like '%cost%';
+-----------------+--------------+
| Variable_name   | Value        |
+-----------------+--------------+
| Last_query_cost | 18006.399000 |
+-----------------+--------------+
1 row in set (0.00 sec)
再创建mobile上的索引:
MariaDB [db_vip]> create index i_m_db_vip_2 on m_db_vip(mobile); 
Query OK, 0 rows affected (0.96 sec)
Records: 0  Duplicates: 0  Warnings: 0


MariaDB [db_vip]> explain select            ext_id, mid, msource_id, msource_type, referee, mobile, tel, mpassword,status         from m_db_vip     where  (msource_id='xx1391434680574433' and msource_type=1 )     or ( mobile='1391434680574433'   and msource_type=1);
+------+-------------+-----------+-------------+-----------------------------+-----------------------------+---------+------+------+-------------------------------------------------------+
| id   | select_type | table     | type        | possible_keys               | key                         | key_len | ref  | rows | Extra                                                 |
+------+-------------+-----------+-------------+-----------------------------+-----------------------------+---------+------+------+-------------------------------------------------------+
|    1 | SIMPLE      | m_db_vip | index_merge | i_m_db_vip_1,i_m_db_vip_2 | i_m_db_vip_1,i_m_db_vip_2 | 98,99   | NULL |    2 | Using union(i_m_db_vip_1,i_m_db_vip_2); Using where |
+------+-------------+-----------+-------------+-----------------------------+-----------------------------+---------+------+------+-------------------------------------------------------+
1 row in set (0.02 sec)


MariaDB [db_vip]> show status like '%cost%';
+-----------------+----------+
| Variable_name   | Value    |
+-----------------+----------+
| Last_query_cost | 6.826060 |
+-----------------+----------+
1 row in set (0.00 sec)


MariaDB [db_vip]> 
SQL的执行成本已经大幅降低了,经开发已测试,效果相当不错,并发从10几个冲到2000以上。
1 0
原创粉丝点击