SQL Server实现自动循环归档分区数据脚本详解(1/2)

概述

大家应该都知道在很多业务场景下我们需要对一些记录量比较大的表进行分区，同时为了保证性能需要将一些旧的数据进行归档。在分区表很多的情况下如果每一次归档都需要人工干预的话工程量是比较大的而且也容易发生纰漏。接下来分享一个自己编写的自动归档分区数据的脚本，原理是分区表和归档表使用相同的分区方案，循环利用当前的文件组，话不多说了，来一起看看详细的介绍吧。

一、创建测试数据

----01创建文件组
USE [master]
GO
ALTER DATABASE [chenmh] ADD FILEGROUP [Group1]
GO
ALTER DATABASE [chenmh] ADD FILEGROUP [Group2]
GO
ALTER DATABASE [chenmh] ADD FILEGROUP [Group3]
GO
ALTER DATABASE [chenmh] ADD FILEGROUP [Group4]
GO
USE [master]
GO
ALTER DATABASE [chenmh] ADD FILE ( NAME = N'datafile1', FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\DATA\datafile1.ndf' , SIZE = 8192KB , FILEGROWTH = 65536KB ) TO FILEGROUP [Group1]
GO
ALTER DATABASE [chenmh] ADD FILE ( NAME = N'datafile2', FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\DATA\datafile2.ndf' , SIZE = 8192KB , FILEGROWTH = 65536KB ) TO FILEGROUP [Group2]
GO
ALTER DATABASE [chenmh] ADD FILE ( NAME = N'datafile3', FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\DATA\datafile3.ndf' , SIZE = 8192KB , FILEGROWTH = 65536KB ) TO FILEGROUP [Group3]
GO
ALTER DATABASE [chenmh] ADD FILE ( NAME = N'datafile4', FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL12.MSSQLSERVER\MSSQL\DATA\datafile4.ndf' , SIZE = 8192KB , FILEGROWTH = 65536KB ) TO FILEGROUP [Group4]
GO

----02创建分区函数
USE [chenmh]
GO
CREATE PARTITION FUNCTION [Pt_Range](BIGINT) AS RANGE RIGHT FOR VALUES (1000000, 2000000, 3000000)
GO

----03创建分区方案,分区方案对应的文件组数是分区函数指定的数量+1
CREATE PARTITION SCHEME Ps_Range
AS PARTITION Pt_Range
TO (Group1, Group2, Group3, Group4);

---04创建表,指定的分区列的数据类型一定要和分区函数指定的列类型一致。
CREATE TABLE [dbo].[News](
 [id] [bigint] NOT NULL,
 [status] [int] NULL,
 CONSTRAINT [PK_News] PRIMARY KEY CLUSTERED 
(
 [id] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [Ps_Range](id)
) ON [Ps_Range](id)

-----创建归档分区表
CREATE TABLE [dbo].[NewsArchived](
 [id] [bigint] NOT NULL,
 [status] [int] NULL,
 CONSTRAINT [PK_NewsArchived] PRIMARY KEY CLUSTERED 
(
 [id] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [Ps_Range](id)
) ON [Ps_Range](id)

----插入测试数据
DECLARE @id INT 
SET @id=1
WHILE @id<5001000
BEGIN
 INSERT INTO News VALUES(@id,@id%2)
 SET @id=@id+1

END

可以看到当前总共有4个分区，每一个分区定义的范围区间是100万，分区4我故意多插入了200多万的数据来验证自动归档分区。

二、自动归档分区脚本

CREATE PROCEDURE Pro_Partition_AutoArchiveData
(@PartitionTable VARCHAR(300),
@SwitchTable VARCHAR(300)
)
AS
BEGIN
DECLARE @FunName VARCHAR(100),@SchemaName VARCHAR(100),@MaxPartitionValue sql_variant

---根据归档表查找对应的分区方案、分区函数、最小分区数、最大分区范围值
SELECT 
DISTINCT
@FunName=MAX(pf.name),
@SchemaName=MAX(ps.name), 
@MaxPartiti
FROM sys.partitions p inner join sys.indexes i ON p.object_id=i.object_id and p.index_id=i.index_id
inner join sys.partition_schemes ps ON i.data_space_id=ps.data_space_id
inner join sys.destination_data_spaces dds ON ps.data_space_id=dds.partition_scheme_id and dds.destinati
inner join sys.data_spaces ds ON dds.data_space_id=ds.data_space_id
inner join sys.partition_functions pf ON ps.functi
LEFT join sys.partition_range_values prv ON pf.functi AND prv.boundary_id=p.partition_number-pf.boundary_value_on_right
LEFT join sys.partition_parameters pp ON prv.functi and prv.parameter_id=pp.parameter_id
LEFT join sys.types t ON pp.system_type_id=t.system_type_id and pp.user_type_id=t.user_type_id
WHERE OBJECT_NAME(p.OBJECT_ID)=@PartitionTable

DECLARE @MaxId BIGINT,@MinId BIGINT,@Sql NVARCHAR(MAX),@GroupName VARCHAR(100),@MinPartitionNumber INT
SET @Sql= N'SELECT @MaxId=MAX(id),@MinId=Min(id) FROM '+@PartitionTable
EXEC sp_executesql @Sql,N'@MaxId BIGINT out,@MinId BIGINT out',@MaxId OUT,@MinId OUT

SELECT @FunName AS FunName,@SchemaName AS SchemaName,@MaxPartitionValue AS MaxPartitionValue ,@MaxId AS MaxId,@MinId AS MinId

---判断当前表的最大的id是否已经在最大的分区中
IF @MaxId>=@MaxPartitionValue
 BEGIN
 ----归档分区数据,根据表的最小值找到它所属的分区.
 SET @Sql= N'SELECT @MinPartiti
 EXEC sp_executesql @Sql,N'@MinPartitionNumber INT out',@MinPartitionNumber OUT
 SET @Sql=N'ALTER TABLE ' +@PartitionTable+ N' SWITCH PARTITION '+CONVERT(VARCHAR(10),@MinPartitionNumber)+ N' TO ' +@SwitchTable+ N' PARTITION ' +CONVERT(VARCHAR(10),@MinPartitionNumber);
 --PRINT @Sql
 EXEC (@Sql)
 ---修改分区方案，增加新的分区对应的文件组，根据最小的分区id找到对应的文件组。
 SELECT 
 DISTINCT
 @GroupName=ds.name
 FROM sys.partitions p inner join sys.indexes i ON p.object_id=i.object_id and p.index_id=i.index_id
 inner join sys.partition_schemes ps ON i.data_space_id=ps.data_space_id
 inner join sys.destination_data_spaces dds ON ps.data_space_id=dds.partition_scheme_id and dds.destinati
 inner join sys.data_spaces ds ON dds.data_space_id=ds.data_space_id
 inner join sys.partition_functions pf ON ps.functi
 WHERE pf.name=@FunName AND ps.name=@SchemaName AND p.partiti
 SET @Sql=N'ALTER PARTITION SCHEME '+@SchemaName+N' NEXT USED '+@GroupName
 --PRINT @Sql
 EXEC (@Sql)
 ---修改分区函数,增加新的分区,增加新的分区范围值，在现有的最大的值的基础上加100万(需要和现有的分区函数的范围保持一致)
 SET @MaxPartiti
 SET @Sql=N'ALTER PARTITION FUNCTION '+@FunName+N'('+N')'+N' SPLIT RANGE ('+CONVERT(VARCHAR(30),@MaxPartitionValue)+N')'
 --PRINT @Sql
 EXEC (@Sql)

 END

END

三、自动归档分区数据

1.首次测试

EXEC Pro_Partition_AutoArchiveData 'news','NewsArchived';

注意：每调用一次归档一个最小分区的数据。

分区表的News分区1的数据被归档到了NewsArchived表中，且创建了分区5，分区5使用的是已归档的分区1的文件组，达到了循环利用文件组的效果。

2.再调用一次归档分区脚本