MySQL Bug? “SHOW TABLE STATUS”报告导入期间的行数波动

时间:2010-02-03 17:11:31

标签: mysql myisam dml

我将4.1mil记录导入离线系统,对我们数据库的一个子集进行一些分析。当我正在运行导入时,我正在尝试使用以下方法检查其进度:

SHOW TABLE STATUS LIKE 'MailIssueElement'

奇怪的是......在不同的时间,我看到Rows的值更高和更低。我希望它只会上升。以下是输出示例:

mysql> show table status like 'MailIssueElement' \G
*************************** 1. row ***************************
           Name: MailIssueElement
         Engine: MyISAM
        Version: 10
     Row_format: Dynamic
           Rows: 2818307
 Avg_row_length: 120
    Data_length: 338392232
Max_data_length: 281474976710655
   Index_length: 158029824
      Data_free: 0
 Auto_increment: 10248973
    Create_time: 2010-02-03 10:58:41
    Update_time: 2010-02-03 11:04:06
     Check_time: 2010-02-03 10:58:53
      Collation: latin1_swedish_ci
       Checksum: NULL
 Create_options:
        Comment:
1 row in set (0.60 sec)

mysql> show table status like 'MailIssueElement' \G
*************************** 1. row ***************************
           Name: MailIssueElement
         Engine: MyISAM
        Version: 10
     Row_format: Dynamic
           Rows: 1870294
 Avg_row_length: 119
    Data_length: 223251912
Max_data_length: 281474976710655
   Index_length: 107688960
      Data_free: 0
 Auto_increment: 10248973
    Create_time: 2010-02-03 10:58:41
    Update_time: 2010-02-03 11:04:13
     Check_time: 2010-02-03 10:58:53
      Collation: latin1_swedish_ci
       Checksum: NULL
 Create_options:
        Comment:
1 row in set (0.35 sec)

mysql> show table status like 'MailIssueElement' \G
*************************** 1. row ***************************
           Name: MailIssueElement
         Engine: MyISAM
        Version: 10
     Row_format: Dynamic
           Rows: 3074205
 Avg_row_length: 120
    Data_length: 369507112
Max_data_length: 281474976710655
   Index_length: 171537408
      Data_free: 0
 Auto_increment: 10248973
    Create_time: 2010-02-03 10:58:41
    Update_time: 2010-02-03 11:04:36
     Check_time: 2010-02-03 10:58:53
      Collation: latin1_swedish_ci
       Checksum: NULL
 Create_options:
        Comment:
1 row in set (0.01 sec)

mysql> show table status like 'MailIssueElement' \G
*************************** 1. row ***************************
           Name: MailIssueElement
         Engine: MyISAM
        Version: 10
     Row_format: Dynamic
           Rows: 1870294
 Avg_row_length: 119
    Data_length: 223251912
Max_data_length: 281474976710655
   Index_length: 107688960
      Data_free: 0
 Auto_increment: 10248973
    Create_time: 2010-02-03 10:58:41
    Update_time: 2010-02-03 11:04:40
     Check_time: 2010-02-03 10:58:53
      Collation: latin1_swedish_ci
       Checksum: NULL
 Create_options:
        Comment:
1 row in set (0.00 sec)

这种行为有解释吗?有没有更好的方法来检查我的导入进度?

运行以下版本:服务器版本:5.0.32-Debian_7etch11-log Debian蚀刻分布


修改

这是DDL。这是一张MyISAM表:

mysql> show create table MailIssueElement \G
*************************** 1. row ***************************
       Table: MailIssueElement
Create Table: CREATE TABLE `MailIssueElement` (
  `Id` int(11) NOT NULL auto_increment,
  `IssueId` int(11) NOT NULL default '0',
  `Date` datetime NOT NULL default '0000-00-00 00:00:00',
  `Direction` enum('inbound','outbound') NOT NULL default 'inbound',
  `ToAddr` varchar(255) NOT NULL default '',
  `FromAddr` varchar(255) NOT NULL default '',
  `CCAddrs` varchar(255) NOT NULL default '',
  `Subject` text NOT NULL,
  `ParentIssueElementId` int(11) default NULL,
  `ParentIssueElementType` enum('mail','phone') default 'mail',
  `AgentId` int(11) NOT NULL default '0',
  PRIMARY KEY  (`Id`),
  KEY `date_idx` (`Date`),
  KEY `IssueId` (`IssueId`),
  KEY `idx_agent_id` (`AgentId`)
) ENGINE=MyISAM AUTO_INCREMENT=15099881 DEFAULT CHARSET=latin1
1 row in set (0.00 sec)

提前致谢,

-AJ

4 个答案:

答案 0 :(得分:3)

关于行数

  

某些存储引擎(如MyISAM)会存储确切的计数。对于其他存储引擎,例如InnoDB,此值是近似值,并且可能与实际值相差多达40%到50%。

取自MySQL :: MySQL 5.1 Reference Manual - SHOW TABLE STATUS Syntax

答案 1 :(得分:2)

该表必须是InnoDB。

来自SHOW STATUS Doc

<强>行

行数。某些存储引擎(如MyISAM)会存储确切的计数。对于其他存储引擎,例如InnoDB,此值是近似值,并且可能与实际值相差多达40%到50%。在这种情况下,请使用SELECT COUNT(*)来获得准确的计数。

答案 2 :(得分:1)

好的,看起来好像是a bug reported and fixed。猜猜我需要最终从Etch升级到Lenny并获得那些更新的包...感谢您尝试所有。

答案 3 :(得分:0)

如果你使用InnoDB,Rows似乎是通过将“Data_length”除以“Avg_row_length”(波动很大)来计算的。