首页 > 数据库 > SQL Server > 正文

SQL Server表分区的NULL值问题

2024-08-31 00:54:14
字体:
来源:转载
供稿:网友
SQL Server表分区的NULL值问题SQL Server表分区的NULL值问题

SQL Server表分区只支持range分区这一种类型,但是本人觉得已经够用了

虽然MySQL支持四种分区类型:RANGE分区、LIST分区、HASH分区、KEY分区,还支持子分区,但是功能有些欠缺

共同点是MySQL跟SQL Server也有分区对齐的问题,都是水平切分,大家都允许分区列存在NULL值

这次我们测试SQL Server表分区的分区列的NULL值,究竟NULL值是被存放在哪个区间,以前一直没有注意

测试脚本

--1.创建文件组ALTER DATABASE [sss]ADD FILEGROUP [FG_TinyBlog_Id_01]ALTER DATABASE [sss]ADD FILEGROUP [FG_TinyBlog_Id_02]ALTER DATABASE [sss]ADD FILEGROUP [FG_TinyBlog_Id_03]ALTER DATABASE [sss]ADD FILEGROUP [FG_TinyBlog_Id_04]ALTER DATABASE [sss]ADD FILEGROUP [FG_TinyBlog_Id_05]--2.创建文件ALTER DATABASE [sss]ADD FILE(NAME = N'FG_TinyBlog_Id_01_data',FILENAME = N'E:/DataBase/sss/FG_TinyBlog_Id_01_data.ndf',SIZE = 96MB, FILEGROWTH = 24MB )TO FILEGROUP [FG_TinyBlog_Id_01];ALTER DATABASE [sss]ADD FILE(NAME = N'FG_TinyBlog_Id_02_data',FILENAME = N'E:/DataBase/sss/FG_TinyBlog_Id_02_data.ndf',SIZE = 96MB, FILEGROWTH = 24MB )TO FILEGROUP [FG_TinyBlog_Id_02];ALTER DATABASE [sss]ADD FILE(NAME = N'FG_TinyBlog_Id_03_data',FILENAME = N'E:/DataBase/sss/FG_TinyBlog_Id_03_data.ndf',SIZE = 96MB, FILEGROWTH = 24MB )TO FILEGROUP [FG_TinyBlog_Id_03];ALTER DATABASE [sss]ADD FILE(NAME = N'FG_TinyBlog_Id_04_data',FILENAME = N'E:/DataBase/sss/FG_TinyBlog_Id_04_data.ndf',SIZE = 96MB, FILEGROWTH = 24MB )TO FILEGROUP [FG_TinyBlog_Id_04];ALTER DATABASE [sss]ADD FILE(NAME = N'FG_TinyBlog_Id_05_data',FILENAME = N'E:/DataBase/sss/FG_TinyBlog_Id_05_data.ndf',SIZE = 96MB, FILEGROWTH = 24MB )TO FILEGROUP [FG_TinyBlog_Id_05];--3.创建分区函数CREATE PARTITION FUNCTIONFun_TinyBlog_Id(INT) ASRANGE LEFTFOR VALUES(-10,0,1,6)--4.创建分区方案CREATE PARTITION SCHEME[Sch_TinyBlog_Id] aspARTITION [Fun_TinyBlog_Id]TO([FG_TinyBlog_Id_01],[FG_TinyBlog_Id_02],[FG_TinyBlog_Id_03],[FG_TinyBlog_Id_04],[FG_TinyBlog_Id_05])
View Code

插入测试数据

USE [sss]CREATE TABLE TinyBlog(id INT  NULL,NAME VARCHAR(100))  ON [Sch_TinyBlog_Id](id)SELECT * FROM TinyBlog ORDER BY id INSERT INTO [dbo].[TinyBlog]        ( [id], [NAME] )VALUES  ( NULL, -- id - int          '3232'  -- NAME - varchar(100)          )INSERT INTO [dbo].[TinyBlog]        ( [id], [NAME] )VALUES  ( -2, -- id - int          '-2'  -- NAME - varchar(100)          )INSERT INTO [dbo].[TinyBlog]        ( [id], [NAME] )VALUES  ( 66, -- id - int          '66'  -- NAME - varchar(100)          )INSERT INTO [dbo].[TinyBlog]        ( [id], [NAME] )VALUES  ( 0, -- id - int          '0'  -- NAME - varchar(100)          )INSERT INTO [dbo].[TinyBlog]        ( [id], [NAME] )VALUES  ( -30, -- id - int          '-30'  -- NAME - varchar(100)          )

表数据如下

SELECT * FROM TinyBlog ORDER BY id 

分区分布

--查看分区架构文件组分布SELECT  CONVERT(VARCHAR(MAX), ps.name) AS partition_scheme ,        p.partition_number ,        CONVERT(VARCHAR(MAX), ds2.name) AS filegroup ,        CONVERT(VARCHAR(MAX), ISNULL(v.value, ''), 120) AS range_boundary ,        STR(p.rows, 9) AS rowsFROM    sys.indexes i        JOIN sys.partition_schemes ps ON i.data_space_id = ps.data_space_id        JOIN sys.destination_data_spaces dds ON ps.data_space_id = dds.partition_scheme_id        JOIN sys.data_spaces ds2 ON dds.data_space_id = ds2.data_space_id        JOIN sys.partitions p ON dds.destination_id = p.partition_number                                 AND p.object_id = i.object_id                                 AND p.index_id = i.index_id        JOIN sys.partition_functions pf ON ps.function_id = pf.function_id        LEFT JOIN sys.Partition_Range_values v ON pf.function_id = v.function_id                                                  AND v.boundary_id = p.partition_number                                                  - pf.boundary_value_on_rightWHERE   i.object_id = OBJECT_ID('TinyBlog')        AND i.index_id IN ( 0, 1 )ORDER BY p.partition_number--分区区间--SELECT  *  FROM    sys.partition_range_values
View Code

分区情况

--分区情况SELECT  c.* ,        b.[groupname] AS '分区方案对应的文件组名称' ,        d.name '当前分区函数对应的分区方案'FROM    sys.destination_data_spaces AS a        INNER JOIN sysfilegroups AS b ON a.[data_space_id] = b.[groupid]        INNER JOIN ( SELECT $PARTITION.Fun_TinyBlog_Id(id) AS 分区编号 ,                            MIN(id) AS Min_value ,                            MAX(id) AS Max_value ,                            COUNT(id) AS 记录数                     FROM   [dbo].[TinyBlog]                     GROUP BY $PARTITION.Fun_TinyBlog_Id(id)                   ) AS c ON c.[分区编号] = a.[destination_id]        INNER JOIN sys.partition_schemes AS d ON a.[partition_scheme_id] = d.data_space_idORDER BY c.[分区编号]
View Code

第二个视图直接把NULL值忽略了

根据第一个视图,我们画出一个总结图

分析一下


总结

结论其实很清晰,分区列不要允许NULL,如果允许列,那么我们交换分区归档数据的时候就会有麻烦了,因为无法判断分区列是NULL值的那些数据行是老数据还是新数据

相关文章:

http://www.cnblogs.com/gaizai/archive/2010/11/05/1870071.html

https://msdn.microsoft.com/zh-cn/library/ms187802.aspx

如有不对的地方,欢迎大家拍砖o(∩_∩)o


发表评论 共有条评论
用户名: 密码:
验证码: 匿名发表