• ADADADADAD

    MySQL中SQL语句优化的示例分析[ mysql数据库 ]

    mysql数据库 时间:2024-11-26 22:14:50

    作者:文/会员上传

    简介:

    sql如下:sql强制用了into_time索引# Time: 2017-02-14T11:35:01.594499+08:00# User@Host: oms_readonly[oms_readonly] @ [10.44.xxx.xxx] Id: 41636892# Query_time: 9.299

    以下为本文的正文内容,内容仅供参考!本站为公益性网站,复制本文以及下载DOC文档全部免费。

    sql如下:sql强制用了into_time索引

      # Time: 2017-02-14T11:35:01.594499+08:00

      # User@Host: oms_readonly[oms_readonly] @ [10.44.xxx.xxx] Id: 41636892

      # Query_time: 9.299612 Lock_time: 0.000124 Rows_sent: 20 Rows_examined: 2439330

      SET timestamp=1487043301;

      select * from customers force index(`into_time`) where `type` = 1 AND `status` < 7 AND `isarea` = 8 AND `into_time`<='2017-01-31 23:59:59' order by score desc limit 0,20;

    查看表的相关状态:

      mysql> show table status like 'customers' \G;

      *************************** 1. row ***************************

      Name: customers

      Engine: InnoDB

      Version: 10

      Row_format: Dynamic

      Rows: 2504609

      Avg_row_length: 710

      Data_length: 1780383744

      Max_data_length: 0

      Index_length: 1253048320

      Data_free: 6291456

      Auto_increment: 2546101

      Create_time: 2017-01-07 01:59:34

      Update_time: 2017-02-14 13:58:17

      Check_time: NULL

      Collation: utf8_general_ci

      Checksum: NULL

      Create_options:

      Comment:

      1 row in set (0.00 sec)

    表一共大约有250万行记录,查看下满足into_time<='2017-01-31 23:59:59'这个条件的有多少行

      mysql> select count(*) from customers where `into_time`<='2017-01-31 23:59:59';

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

      | count(*) |

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

      | 2439147 |

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

      1 row in set (0.95 sec)


    显然into_time这个列的索引已经不合适了,查看下表上都有那些索引

      Create Table: CREATE TABLE `customers` (

      `id` int(11) NOT NULL AUTO_INCREMENT,

      PRIMARY KEY (`id`),

      KEY `newdata` (`newdata`),

      KEY `cusname` (`cusname`),

      KEY `type` (`type`,`ownerid`),

      KEY `operator` (`operator`),

      KEY `into_time` (`into_time`),

      KEY `isarea` (`isarea`),

      KEY `linkcase` (`linkcase`),

      KEY `score` (`score`),

      FULLTEXT KEY `fdx_cusname` (`cusname_idx`)

      ) ENGINE=InnoDB AUTO_INCREMENT=2546101 DEFAULT CHARSET=utf8

      1 row in set (0.00 sec)

    可以看到score列有索引,如果能采用这个列的索引是个比较好的选择,去掉强制索引看下执行计划



        mysql> explain select * from customers where `type` = 1 AND `status` < 7 AND `isarea` = 8 AND `into_time`<='2017-01-31 23:59:59' order by score desc limit 0,20;

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

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

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

        | 1 | SIMPLE | customers | NULL | index | type,into_time,isarea,status,idx_isarea_renew_owner,type_status | score | 2 | NULL | 270 | 0.92 | Using where |

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

        1 row in set, 1 warning (0.00 sec)

    可以看到用了score索引,执行时间从最9秒多,到优化后的0.0几秒。

    MySQL中SQL语句优化的示例分析.docx

    将本文的Word文档下载到电脑

    推荐度:

    下载
    热门标签: mysqlsql