MySQL中distinct和group by性能比较[转]

MySQL中distinct和group by性能比较[转]

之前看了网上的一些测试，感觉不是很准确，今天亲自测试了一番。得出了结论(仅在个人计算机上测试，可能不全面，仅供参考)

测试过程：

准备一张测试表

1    CREATE TABLE `test_test` (
2      `id` int(11) NOT NULL auto_increment,
3      `num` int(11) NOT NULL default '0',
4      PRIMARY KEY  (`id`)
5     ) ENGINE=MyISAM  DEFAULT CHARSET=utf8 AUTO_INCREMENT=1 ;

建个储存过程向表中插入10W条数据

01    create procedure p_test(pa int(11))
02     begin
03
04      declare max_num int(11) default 100000;
05      declare i int default 0;
06      declare rand_num int;
07
08      select count(id) into max_num from test_test;
09
10      while i < pa do
11              if max_num < 100000 then
12                      select cast(rand()*100 as unsigned) into rand_num;
13                      insert into test_test(num)values(rand_num);
14              end if;
15              set i = i +1;
16      end while;
17     end

调用存储过程插入数据

1    call p_test(100000);

开始测试：（不加索引）

01    select distinct num from test_test;
02    select num from test_test group by num;
03
04    [SQL] select distinct num from test_test;
05    受影响的行: 0
06    时间: 0.078ms
07
08    [SQL]
09    select num from test_test group by num;
10    受影响的行: 0
11    时间: 0.031ms

二、num字段上创建索引

1    ALTER TABLE `test_test` ADD INDEX `num_index` (`num`) ;

再次查询

01    select distinct num from test_test;
02    select num from test_test group by num;
03    [SQL] select distinct num from test_test;
04    受影响的行: 0
05    时间: 0.000ms
06
07    [SQL]
08    select num from test_test group by num;
09    受影响的行: 0
10    时间: 0.000ms

这时候我们发现时间太小了 0.000秒都无法精确了。

我们转到命令行下 测试

01    mysql> set profiling=1;
02    mysql> select distinct(num) from test_test;
03    mysql> select num from test_test group by num;
04    mysql> show profiles;
05    +----------+------------+----------------------------------------+
06    | Query_ID | Duration   | Query                                  |
07    +----------+------------+----------------------------------------+
08    |        1 | 0.00072550 | select distinct(num) from test_test    |
09    |        2 | 0.00071650 | select num from test_test group by num |
10    +----------+------------+----------------------------------------+

加了索引之后 distinct 比没加索引的 distinct 快了 107倍。

加了索引之后 group by 比没加索引的 group by 快了 43倍。

再来对比 ：distinct  和 group by

不管是加不加索引 group by 都比 distinct 快。因此使用的时候建议选 group by

来源： <http://www.cnblogs.com/zox2011/archive/2012/09/12/2681797.html>

来自为知笔记(Wiz)