zoukankan      html  css  js  c++  java
  • Mysql-SQL优化-统计某种类型的个数

    有时我们想统计某种类型有多少个,会用这个SQL。

    全表扫描之余,还要filesort。耗时1.34秒。

     
    mysql>  select country,count(*) from t1 group by country;
    +---------+----------+
    | country | count(*) |
    +---------+----------+
    | NULL    |       32 |
    | africa  |   524288 |
    | america |   524288 |
    | china   |   524288 |
    +---------+----------+
    4 rows in set (1.34 sec)
    
    mysql> desc select country,count(*) from t1 group by country;
    +----+-------------+-------+------+---------------+------+---------+------+---------+---------------------------------+
    | id | select_type | table | type | possible_keys | key  | key_len | ref  | rows    | Extra                           |
    +----+-------------+-------+------+---------------+------+---------+------+---------+---------------------------------+
    |  1 | SIMPLE      | t1    | ALL  | NULL          | NULL | NULL    | NULL | 1573382 | Using temporary; Using filesort |
    +----+-------------+-------+------+---------------+------+---------+------+---------+---------------------------------+
    1 row in set (0.00 sec)
    
    
    下面是两种优化方法,都是全表扫描,但预计count()要比sum()耗的CPU少点,方法更佳。

    mysql> select count(country='africa' or null) as africa,count(country='america' or null) as america, count(country='china' or null) as china from t1;
    +--------+---------+--------+
    | africa | america | china  |
    +--------+---------+--------+
    | 524288 |  524288 | 524288 |
    +--------+---------+--------+
    1 row in set (0.78 sec)
    
    mysql> desc select count(country='africa' or null) as africa,count(country='america' or null) as america, count(country='china' or null) as china from t1;
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    | id | select_type | table | type | possible_keys | key  | key_len | ref  | rows    | Extra |
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    |  1 | SIMPLE      | t1    | ALL  | NULL          | NULL | NULL    | NULL | 1573382 |       |
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    1 row in set (0.00 sec)
    
    
    
    mysql> select sum(country='africa') as africa ,sum(country='america') as america,sum(country='china') from t1;
    +--------+---------+----------------------+
    | africa | america | sum(country='china') |
    +--------+---------+----------------------+
    | 524288 |  524288 |               524288 |
    +--------+---------+----------------------+
    1 row in set (0.86 sec)
    
    mysql> desc select sum(country='africa') as africa ,sum(country='america') as america,sum(country='china') from t1;
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    | id | select_type | table | type | possible_keys | key  | key_len | ref  | rows    | Extra |
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    |  1 | SIMPLE      | t1    | ALL  | NULL          | NULL | NULL    | NULL | 1573382 |       |
    +----+-------------+-------+------+---------------+------+---------+------+---------+-------+
    1 row in set (0.00 sec)
    
    
    版本号mysql5.5.30


  • 相关阅读:
    hdu 3579 Hello Kiki (中国剩余定理)
    hdu 1573 X问题 (非互质的中国剩余定理)
    poj 2891 Strange Way to Express Integers (非互质的中国剩余定理)
    查看cache中消耗性能的语句
    sqlserver2008 R2中查找未使用过的索引
    【转】如何新建索引:高开销的缺失索引
    查看sqlserver 2008中性能低下的语句
    经典
    符合条件的记录有则修改没有则添加的小优化技巧
    测试分区表,以及在现有表上新建分区方案,以及把分区表变为普通表测试代码!!
  • 原文地址:https://www.cnblogs.com/claireyuancy/p/6855142.html
Copyright © 2011-2022 走看看