sql server表分区实验

--***********分区表实验*******************
--*        Write by yuch_sheng on 2020-07-05   *
--**************************************

--新建文件组
USE master
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUPOLD]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202001]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202002]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202003]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202004]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202005]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202006]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202007]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202008]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202009]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202010]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202011]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP202012]
GO
--查看文件组信息
sp_helpfilegroup
-- 向文件组中添加文件
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE (NAME  = N'FILEOLD', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_OLD.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUPOLD]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202001', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202001.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202001]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202002', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202002.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202002]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202003', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202003.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202003]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202004', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202004.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202004]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202005', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202005.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202005]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202006', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202006.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202006]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202007', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202007.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202007]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202008', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202008.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202008]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202009', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202009.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202009]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202010', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202010.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202010]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202011', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202011.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202011]
GO
ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE202012', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_202012.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP202012]
GO

--创建分区函数
CREATE PARTITION FUNCTION [PFUN_DATETIME](datetime) AS RANGE RIGHT FOR VALUES 
(
   N'2020-01-01T00:00:00', 
   N'2020-02-01T00:00:00', 
   N'2020-03-01T00:00:00', 
   N'2020-04-01T00:00:00', 
   N'2020-05-01T00:00:00', 
   N'2020-06-01T00:00:00', 
   N'2020-07-01T00:00:00', 
   N'2020-08-01T00:00:00', 
   N'2020-09-01T00:00:00', 
   N'2020-10-01T00:00:00', 
   N'2020-11-01T00:00:00', 
   N'2020-12-01T00:00:00'
 );
 GO

--创建分区方案
CREATE PARTITION SCHEME [PS_PFUN_DATETIME]
       AS PARTITION [PFUN_DATETIME]
       TO
        (
           [FILEGROUPOLD], 
           [FILEGROUP202001],
           [FILEGROUP202002],
           [FILEGROUP202003],
           [FILEGROUP202004],
           [FILEGROUP202005],
           [FILEGROUP202006],
           [FILEGROUP202007],
           [FILEGROUP202008],
           [FILEGROUP202009],
           [FILEGROUP202010],
           [FILEGROUP202011],
           [FILEGROUP202012],
           [PRIMARY]
         );

    -- 针对分区列创建聚集索引,
    -- 此时表中的数据会按分区方案的定义重排在不同的文件组中
        USE [MES_WORKBECH_TD]
        GO

        CREATE CLUSTERED INDEX [cidx_createDT] ON [dbo].[PRODUCT_WORKORDER_PROCESS_PU]
        (
            [CreateDT]
        )
        WITH (SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = OFF)
           ON [PS_PFUN_DATETIME]([CreateDT])

   --查看每给分区的记录数
       select $PARTITION.PFUN_DATETIME(CreateDT) as 分区编号,
                count(ID) as 记录数 
        from [PRODUCT_WORKORDER_PROCESS_PU] 
        group by $PARTITION.PFUN_DATETIME(CreateDT) 
  
  --分区结合index提升系统性能
      select  count(1)
         from  [PRODUCT_WORKORDER_PROCESS_PU] a
       where 
           a.CreateDT > '2020-05-25 11:53:32.093'  --产线计划的建单时间
           and a.LinePlanId = '2E03C2E1-3809-47EC-9014-21B30B9C532D'
           and a.ProcessCode = 'PU_P003'
           and a.State = '1'
           and a.Result = '1'
   
        select  *
         from  [PRODUCT_WORKORDER_PROCESS_PU] a
       where 
           a.CreateDT > '2020-05-25 11:53:32.093'
           and a.ProcessCode = 'PU_P003'
           and a.TrackInBarCode = '5473347I919C520221SA1014245'
           --and a.TrackOutBarCode = '5473347I919C520221SA1014245'
 

        drop index [idx_product_workorder_process_pu_1] 
          on [PRODUCT_WORKORDER_PROCESS_PU];
  
         create index [idx_product_workorder_process_pu_1]
         on [PRODUCT_WORKORDER_PROCESS_PU]
         (
            LinePlanId,ProcessCode,State,Result
         )

         create index [idx_product_workorder_process_pu_2]
         on [PRODUCT_WORKORDER_PROCESS_PU]
         (
            TrackInBarCode,ProcessCode
         )
         include(id);

         create index [idx_product_workorder_process_pu_3]
         on [PRODUCT_WORKORDER_PROCESS_PU]
         (
            TrackOutBarCode,ProcessCode
         )
         include(id)

     --*****************************************
     --*                 分区的合并与新增                            *
     --*****************************************
     --新建第一季度2020Q1文件组
       ALTER DATABASE [MES_WORKBECH_TD] ADD FILEGROUP [FILEGROUP2020Q1]
       GO
     --向2020Q1文件组中添加文件
       ALTER DATABASE [MES_WORKBECH_TD] ADD FILE ( NAME = N'FILE2020Q1', FILENAME = N'C:dataSQLDB_DATAMES_WORKBECH_2020Q1.mdf' , SIZE = 1048576KB , FILEGROWTH = 1024KB ) TO FILEGROUP [FILEGROUP2020Q1]
       GO
     --合并分区
       ALTER PARTITION FUNCTION PFUN_DATETIME() MERGE RANGE (N'2020-01-01T00:00:00');
       ALTER PARTITION FUNCTION PFUN_DATETIME() MERGE RANGE (N'2020-02-01T00:00:00');  
       ALTER PARTITION FUNCTION PFUN_DATETIME() MERGE RANGE (N'2020-03-01T00:00:00');  
     --观察各分区记录数的变化,可以发现分区数减少了，原分区2,3,4已被合并
       select $PARTITION.PFUN_DATETIME(CreateDT) as 分区编号,
               count(ID) as 记录数 
        from [PRODUCT_WORKORDER_PROCESS_PU] 
        group by $PARTITION.PFUN_DATETIME(CreateDT) 
     --分区新增,将2020年第一季度的数据放入FILEGROUP2020Q1文件组
     --选择文件组
        ALTER PARTITION SCHEME PS_PFUN_DATETIME  
        NEXT USED [FILEGROUP2020Q1] ;
     --修改分区函数  
       ALTER PARTITION FUNCTION PFUN_DATETIME()  
       SPLIT RANGE (N'2020-01-01T00:00:00.000') ;
    --观察各分区记录数的变化,可以发现多了一个分区
       select $PARTITION.PFUN_DATETIME(CreateDT) as 分区编号,
               count(ID) as 记录数 
        from [PRODUCT_WORKORDER_PROCESS_PU] 
        group by $PARTITION.PFUN_DATETIME(CreateDT) 
     --如果分区变动比较大不推荐用合并和删除的方法，因为容易出错，
     --如果分12个月建议像下面一样，先将分区表转换为普通表，再把普通表分区。
     --  a)删除分区索引,删除分区索引后，并没有变成普通表
          drop index [cidx_createDT] on [dbo].[PRODUCT_WORKORDER_PROCESS_PU];
     --  b)在原有分区索引字段，建立普通索引
         CREATE CLUSTERED INDEX [cidx_createDT] ON [dbo].[PRODUCT_WORKORDER_PROCESS_PU]
         (
            [CreateDT]
         )
         ON[PRIMARY];
     -- c) 删除普通萦引
         drop index [cidx_createDT] on [dbo].[PRODUCT_WORKORDER_PROCESS_PU];
     -- d) 再将普通表做分区
        CREATE CLUSTERED INDEX [cidx_createDT] ON [dbo].[PRODUCT_WORKORDER_PROCESS_PU]
        (
            [CreateDT]
        )
        WITH (SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = OFF)
           ON [PS_PFUN_DATETIME]([CreateDT])