【转】利用optimize、存储过程和系统表对mysql数据库表进行批量碎片清理释放表空间

标签：

本文收集于本人的笔记本，由于找不到原文出处。在此省略，如哪位知道可以联系我加上。

核心是利用mysql系统表和“optimize table 表名”命令，对mysql数据表进行空间的释放。由于delete和drop table都不会释放表空间（truncate 命令会释放表空间【将所有的数据都删除】），所以需要利用optimize 命令进行释放。

这个存储过程目的是给一个库的所有表来整理碎片的。一个表随着插入很频繁，或者一直更新不停的，就会积累好多碎片。如果及时整理一下，查询效率会高出好多。

DELIMITER $$

DROP PROCEDURE IF EXISTS `mysql`.`sp_optimize_tables`$$

CREATE PROCEDURE `mysql`.`sp_optimize_tables`(
 IN db_name varchar(255))
BEGIN
  -- Created by david yeung 20080128.

  -- To optimize all the tables in exact database.

  declare cnt int default 0;
  declare i int default 0;
  select count(*) as total from information_schema.tables where table_schema = db_name into cnt;
  while i < cnt
  do
    -- Get the table‘s exact name.
    set @stmt = concat(‘select table_name from information_schema.tables where table_schema = ‘‘‘,db_name,‘‘‘ order by table_name asc limit ‘,i,‘,1 into @tb_name‘);
    prepare s1 from @stmt;
    execute s1;
    drop prepare s1;
    set @stmt = ‘‘;
    set @stmt = concat(‘optimize table ‘,db_name,‘.‘,@tb_name);
    prepare s1 from @stmt;
    execute s1;
    drop prepare s1;
    set @stmt = ‘‘;
    set i = i + 1;
  end while;
  -- Refresh tables.

  flush tables;
END$$

DELIMITER ;

调用示例：

mysql> use mysql
Database changed
mysql> call sp_optimize_tables(‘david_test‘);
+------------------------------+----------+----------+----------+
| Table                        | Op       | Msg_type | Msg_text |
+------------------------------+----------+----------+----------+
| david_test.test1 | optimize | status   | OK       |
+------------------------------+----------+----------+----------+
1 row in set (0.26 sec)

+--------------------------+----------+----------+----------+
| Table                    | Op       | Msg_type | Msg_text |
+--------------------------+----------+----------+----------+
| david_test.test2| optimize | status   | OK       |
+--------------------------+----------+----------+----------+
1 row in set (0.35 sec)

+---------------------------------------+----------+----------+----------+
| Table                                 | Op       | Msg_type | Msg_text |
+---------------------------------------+----------+----------+----------+
| david_test.test3 | optimize | status   | OK       |
+---------------------------------------+----------+----------+----------+
1 row in set (0.45 sec)

+--------------------------+----------+----------+----------+
| Table                    | Op       | Msg_type | Msg_text |
+--------------------------+----------+----------+----------+
| david_test.test_article | optimize | status   | OK       |
+--------------------------+----------+----------+----------+
1 row in set (4.13 sec)

...

+----------------------------------+----------+----------+----------+
| Table                            | Op       | Msg_type | Msg_text |
+----------------------------------+----------+----------+----------+
| david_test.test_article_content | optimize | status   | OK       |
+----------------------------------+----------+----------+----------+
1 row in set (37.81 sec)

+-----------------------+----------+----------+----------+
| Table                 | Op       | Msg_type | Msg_text |
+-----------------------+----------+----------+----------+
| david_test.members | optimize | status   | OK       |
+-----------------------+----------+----------+----------+
1 row in set (40.02 sec)

+--------------------+----------+----------+----------+
| Table              | Op       | Msg_type | Msg_text |
+--------------------+----------+----------+----------+
| david_test.test_site | optimize | status   | OK       |
+--------------------+----------+----------+----------+
1 row in set (40.31 sec)

+--------------+----------+----------+----------+
| Table        | Op       | Msg_type | Msg_text |
+--------------+----------+----------+----------+
| david_test.t | optimize | status   | OK       |
+--------------+----------+----------+----------+
1 row in set (41.10 sec)

Query OK, 0 rows affected (41.13 sec)

【转】利用optimize、存储过程和系统表对mysql数据库表进行批量碎片清理释放表空间

标签：

原文地址：http://www.cnblogs.com/rwxwsblog/p/4528577.html