如何在没有任何渐进字段的情况下对 table 进行分区？

Question

我需要对 table 进行分区以获取值范围内的记录。

我在table中没有渐进字段。

select [_Load_Timestamp], [COL_A], [COL_B], [COL_C], [COL_D], [COL_E], [COL_F]
from dbo.table
where ROW_NUMBER() OVER (order by [_Load_Timestamp], [COL_A], [COL_B], [COL_C]) between 500 and 1000;

编写这段代码时，它 returns 我出现了以下错误：

Windowed functions can only appear in the SELECT or ORDER BY clauses.

问题是 ROW_NUMBER() 语句。

我可以这样写查询，但问题是我必须尽可能地限制对 table 的访问，因为它真的很慢。

select *
from ( 
        select ROW_NUMBER() OVER (order by [_Load_Timestamp], [COL_A], [COL_B], [COL_C])) [id]
                ,[_Load_Timestamp], [COL_A], [COL_B], [COL_C], [COL_D], [COL_E], [COL_F]
        from dbo.table
     ) t
where t.id between 500 and 1000;

谁能给我一个达到相同目标的替代方案？

Answer 1

试试这个：

;with MyTable as  
    ( 
        select 
                ROW_NUMBER() OVER (order by [_Load_Timestamp], [COL_A], [COL_B], [COL_C])) [id]
                ,[_Load_Timestamp], [COL_A], [COL_B], [COL_C], [COL_D], [COL_E], [COL_F]
        from dbo.table
     )
select  *
from MyTable
where id between 500 and 1000
;

Answer 2

我找到了解决方案。

select  ROW_NUMBER() OVER (order by [_Load_Timestamp], [COL_A], [COL_B], [COL_C])) [id], [_Load_Timestamp], [COL_A], [COL_B], [COL_C]
into #tmp_key
from dbo.table

create index IX_01 on #tmp_key on ([id])

select _b.*
from #tmp_key _a
join dbo.table _b on _a.id between 500 and 1000
and _a.[_Load_Timestamp] = _b.[_Load_Timestamp] 
and _a.[COL_A] = _b.[COL_A]
and _a...

数字 500 和 1000 将根据为我填充 table 的循环动态设置。

无论如何，我的问题都解决了。

如何在没有任何渐进字段的情况下对 table 进行分区？

how to partition a table without any progressive fields?

tsql

sql-server

partitioning

row-number