SQL Server 2005:查看分区表(partition table)的分区范围(partition range)

SQL Server 2005 的分区表(partition table)是复杂的,特别是对于初学者来说。不管怎样,我们还是掌握了分区函数(partition function),分区方案(partition scheme),最后终于创建了一个分区表出来。但之后呢,或许你想查看分区表的各个分区分区列的取值范围(这个分区的理论最小值和最大值),虽然可以通过分区函数中定义的边界值(boundary value)来推算出来分区表每个分区(partition)的取值范围,但是事情并不是很简单,你需要一系列繁琐的步骤来实现:首先要找出这个分区表的分区函数;然后确定该分区函数定义的边界值属于 left 或者 right; 最后在脑子里使劲思索几下终于确定了分区列取值范围(partition range)。幸运的是,有了下面的这个存储过程,你的大脑就可以获得解放了。

if exists (select 1 from sys.procedures where name = 'sp_show_partition_range')
   drop procedure dbo.sp_show_partition_range
go

--------------------------------------------------------------------------------
-- author : p.c.w.l
-- source : www.sqlstudy.com
-- create : 2008-01-01
-- descr  : view partition range by 'partition table' or 'partition function'
--------------------------------------------------------------------------------

create procedure dbo.sp_show_partition_range
(
   @partition_table    nvarchar(255) = null
  ,@partition_function nvarchar(255) = null
)
as
begin
   set nocount on

   declare @function_id int
       set @function_id = null

   -- get @function_id base on @partition_table
   if len(@partition_table) > 0 begin
      select @function_id = s.function_id
        from sys.indexes i
                inner join sys.partition_schemes s
          on i.data_space_id = s.data_space_id
       where i.index_id < 2
         and i.object_id = object_id(@partition_table)
         
      if @function_id is null
         return 1         
   end

   -- get @function_id base on @partition_function
   if len(@partition_function) > 0 begin
      select @function_id = function_id
        from sys.partition_functions
       where name = @partition_function
       
      if @function_id is null
         return 1        
    end
   

   -- get partition range
   select partition_function = f.name
         ,t.partition
         ,t.minval
         ,value = case when f.boundary_value_on_right=1 then '<= val <' else '< val <=' end
         ,t.maxval
     from (
           select h.function_id
                 ,partition = h.boundary_id
                 ,minval    = l.value
                 ,maxval    = h.value
             from sys.partition_range_values h
                     left join sys.partition_range_values l
               on h.function_id = l.function_id and h.boundary_id = l.boundary_id + 1

           union all

           select function_id
                 ,partition = max(boundary_id) + 1
                 ,minval    = max(value)
                 ,maxval    = null
             from sys.partition_range_values
            group by function_id
          ) t
              inner join sys.partition_functions f
         on t.function_id = f.function_id
      where f.function_id = @function_id
         or @function_id is null
      order by 1, 2
end
go
原文地址:https://www.cnblogs.com/chuncn/p/1395165.html