SQL將varchar範圍拆分爲單個數字

Question

如果我在某列中有一個範圍，例如

1-5,6,7 

我想輸出：

1 
2 
3 
4 
5 
6 
7 

這是可能的SQL？我正在使用SQL服務器管理工​​作室。

Answer 1

首先我們分裂的字符串，並創建R1和R2，然後我們採用一個特設理貨表以填補這一空白

Declare @YourTable table (ID int,RangeString varchar(max)) 
Insert Into @YourTable values 
(1,'1-5,6,7'), 
(2,'25000-25010,10-15') 


Select A.ID 
     ,C.N 
From @YourTable A 
Cross Apply (
     Select R1=cast(left(RetVal,charindex('-',RetVal+'-')-1) as int) 
      ,R2=cast(case when charindex('-',RetVal)=0 then left(RetVal,charindex('-',RetVal+'-')-1) else replace(RetVal,left(RetVal,charindex('-',RetVal+'-')),'') end as int) 
     From (
       Select RetSeq = Row_Number() over (Order By (Select null)) 
         ,RetVal = LTrim(RTrim(B.i.value('(./text())[1]', 'varchar(max)'))) 
       From (Select x = Cast('<x>' + replace((Select replace(RangeString,',','§§Split§§') as [*] For XML Path('')),'§§Split§§','</x><x>')+'</x>' as xml).query('.')) as X 
       Cross Apply x.nodes('x') AS B(i) 
      ) P 
    ) B 
Cross Apply ( 
       Select Top (B.R2-B.R1+1) N=Row_Number() Over (Order By (Select null))+B.R1-1 From master..spt_values N1,master..spt_values N2 
      ) C 

返回

ID N 
1 1 
1 2 
1 3 
1 4 
1 5 
1 6 
1 7 
2 25000 
2 25001 
2 25002 
2 25003 
2 25004 
2 25005 
2 25006 
2 25007 
2 25008 
2 25009 
2 25010 
2 10 
2 11 
2 12 
2 13 
2 14 
2 15 

編輯 - 使用UDF的清潔版

Select A.ID 
     ,D.N 
From @YourTable A 
Cross Apply [dbo].[udf-Str-Parse](A.RangeString,',') B 
Cross Apply (
       Select R1=min(cast(RetVal as int)) 
        ,R2=max(cast(RetVal as int)) 
       From [dbo].[udf-Str-Parse](B.RetVal,'-') 
      ) C 
Cross Apply ( 
       Select Top (C.R2-C.R1+1) N=Row_Number() Over (Order By (Select null))+C.R1-1 
        From master..spt_values N1,master..spt_values N2 
      ) D 

的UDF如果有興趣

CREATE FUNCTION [dbo].[udf-Str-Parse] (@String varchar(max),@Delimiter varchar(10)) 
Returns Table 
As 
Return ( 
    Select RetSeq = Row_Number() over (Order By (Select null)) 
      ,RetVal = LTrim(RTrim(B.i.value('(./text())[1]', 'varchar(max)'))) 
    From (Select x = Cast('<x>' + replace((Select replace(@String,@Delimiter,'§§Split§§') as [*] For XML Path('')),'§§Split§§','</x><x>')+'</x>' as xml).query('.')) as A 
    Cross Apply x.nodes('x') AS B(i) 
); 
--Thanks Shnugo for making this XML safe 
--Select * from [dbo].[udf-Str-Parse]('Dog,Cat,House,Car',',') 
--Select * from [dbo].[udf-Str-Parse]('John Cappelletti was here',' ') 
--Select * from [dbo].[udf-Str-Parse]('this,is,<test>,for,< & >',',') 
--Performance On a 5,000 random sample -8K 77.8ms, -1M 79ms (+1.16), -- 91.66ms (+13.8)