转 http://blog.csdn.net/m582445672/article/details/7800694 1.创建一个测试表 [sql] view plaincopyprint? CREATE TABLE test( id VARCHAR (20) NOT NULL , name VARCHAR (20) NOT NULL , submit_timeDATETIME NOT NULL , index time_index(submit_time),

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

1.创建一个测试表

[sql] view plaincopyprint?

1. CREATE TABLE test (
2. id VARCHAR(20) NOT NULL,
3. name VARCHAR(20) NOT NULL,
4. submit_time DATETIME NOT NULL,
5. index time_index (submit_time),
6. index id_index (id)
7. )ENGINE=MyISAM
8. PARTITION BY RANGE COLUMNS(submit_time)
9. (
10. PARTITION p1 VALUES LESS THAN ('2010-02-01'),
11. PARTITION p2 VALUES LESS THAN ('2010-03-01'),
12. PARTITION p3 VALUES LESS THAN ('2010-04-01'),
13. PARTITION p4 VALUES LESS THAN ('2010-05-01'),
14. PARTITION p5 VALUES LESS THAN ('2010-06-01'),
15. PARTITION p6 VALUES LESS THAN ('2010-07-01'),
16. PARTITION p7 VALUES LESS THAN ('2010-08-01'),
17. PARTITION p8 VALUES LESS THAN ('2010-09-01'),
18. PARTITION p9 VALUES LESS THAN ('2010-10-01'),
19. PARTITION p10 VALUES LESS THAN ('2010-11-01'),
20. PARTITION p11 VALUES LESS THAN ('2010-12-01')
21. );

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

[sql] view plaincopyprint?

1. delimiter //
2. CREATE PROCEDURE mark_test()
3. begin
4. declare v int default 0;
5. while v < 8000
6. do
7. insert into test values (v,'testing partitions',adddate('2010-01-01', INTERVAL v hour));
8. set v = v + 1;
9. end while;
10. end //
11. 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