MySQL 5.5 分区性能测试之索引使用情况

转 http://blog.csdn.net/m582445672/article/details/7800694

1.创建一个测试表

CREATE TABLE test (
  id VARCHAR(20) NOT NULL,
  name VARCHAR(20) NOT NULL,
  submit_time DATETIME NOT NULL,
  index time_index (submit_time),
  index id_index (id)
)ENGINE=MyISAM
PARTITION BY RANGE COLUMNS(submit_time)
(
PARTITION p1 VALUES LESS THAN ('2010-02-01'),
PARTITION p2 VALUES LESS THAN ('2010-03-01'),
PARTITION p3 VALUES LESS THAN ('2010-04-01'),
PARTITION p4 VALUES LESS THAN ('2010-05-01'),
PARTITION p5 VALUES LESS THAN ('2010-06-01'),
PARTITION p6 VALUES LESS THAN ('2010-07-01'),
PARTITION p7 VALUES LESS THAN ('2010-08-01'),
PARTITION p8 VALUES LESS THAN ('2010-09-01'),
PARTITION p9 VALUES LESS THAN ('2010-10-01'),
PARTITION p10 VALUES LESS THAN ('2010-11-01'),
PARTITION p11 VALUES LESS THAN ('2010-12-01')
);

CREATE TABLE test ( 
  id VARCHAR(20) NOT NULL,
  name VARCHAR(20) NOT NULL,
  submit_time DATETIME NOT NULL,
  index time_index (submit_time),
  index id_index (id)
)ENGINE=MyISAM
PARTITION BY RANGE COLUMNS(submit_time)
(
PARTITION p1 VALUES LESS THAN ('2010-02-01'),
PARTITION p2 VALUES LESS THAN ('2010-03-01'),
PARTITION p3 VALUES LESS THAN ('2010-04-01'),
PARTITION p4 VALUES LESS THAN ('2010-05-01'),
PARTITION p5 VALUES LESS THAN ('2010-06-01'),
PARTITION p6 VALUES LESS THAN ('2010-07-01'),
PARTITION p7 VALUES LESS THAN ('2010-08-01'),
PARTITION p8 VALUES LESS THAN ('2010-09-01'),
PARTITION p9 VALUES LESS THAN ('2010-10-01'),
PARTITION p10 VALUES LESS THAN ('2010-11-01'),
PARTITION p11 VALUES LESS THAN ('2010-12-01')  
);

2.写一个存储过程,插入数据

delimiter //
CREATE PROCEDURE mark_test()
begin
    declare v int default 0;
    while v < 8000
    do
       insert into test values (v,'testing partitions',adddate('2010-01-01', INTERVAL v hour));
       set v = v + 1;
    end while;
end //
delimiter ;

delimiter //  
CREATE PROCEDURE mark_test()
begin 
    declare v int default 0; 
    while v < 8000 
    do 
       insert into test values (v,'testing partitions',adddate('2010-01-01', INTERVAL v hour));
       set v = v + 1;
    end while;
end //
delimiter ;

3.实验开始

 上面可以看到,这个是查某一个分区里面的某一些内容,所以完全可以用到index.效果很好..

 

 上面可以看到,跨分区查询,效果也非常不错.

 

 上面可以到看,跨分区查询是,如果某个分区没有用到索引(p4就是全表扫描),整个也没有用到index.但好的是,只扫描需要的分区

 

 

 上面可以看到,如果你不用分区的字段查询,是很杯具的,因为MySQL不知道你分区的index是分别存放到哪个分区上,所以要全index扫描, 

 

3.顺便看看表结构

a. 图中test3 是innodb的存储引擎, 

    test3.frm是表结构.

    test3.par是分区表的信息.

    数据和索引都是存放在表空间里面在

b.图中test是myisam的存储引擎, 

   test.frm是表结构,

   test.par是分区表的信息.

   test#P#p10.MYD是数据文件之一,

   test#P#p10.MYI是索引文件之一