Question

我在SQL Server 2008 db中有大量行

对于每一行，我有3列我关心

典型的行如下所示：

AccountNumber    | basecode                      | subcode
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
9689787209368901 | AQTXG AQTXG AQTXG AQTXG ACC5Z | ZQ596 ZQ596 ZQ596 ZQ655 ZC655

某些基本代码需要与相应的子代码一起删除。 basecode和supcode都是由空格分隔的有序列表。我需要一个查询，我输入基本代码，它删除基本代码和子代码，并保持列表的其余部分完整。

每行具有相同数量的基码和子码项。代码数量可以在0到45之间变化。相同的基本代码可以多次出现。所有字段都是varchars。

这样数据最终（动态解析）如下:(因为每个acct有一个可变数量的基本码/子码（对）。

如果我想删除“ACC5Z”，那么我应该

AccountNumber    | basecode                | subcode
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
9689787209368901 | AQTXG AQTXG AQTXG AQTXG | ZQ596 ZQ596 ZQ596 ZQ655

如果我想删除“AQTXG”，那么我应该

AccountNumber    | basecode   | subcode
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
9689787209368901 | ACC5Z      | ZC655

Answer 1

试试这个

DECLARE @t TABLE (AccountNumber VARCHAR(100),Basecode VARCHAR(100),Subcode VARCHAR(100))
INSERT INTO @t SELECT 9689787209368901,'AQTXG AQTXG AQTXG AQTXG ACC5Z','ZQ596 ZQ596 ZQ596 ZQ655 ZC655'

DECLARE @BaseCodetoRemove VARCHAR(10) = 'AQTXG'

;WITH CTE AS(
SELECT  
    AccountNumber
    ,LTRIM(RTRIM(SUBSTRING(Basecode, Number ,CHARINDEX(' ', Basecode + ' ', Number ) - Number))) AS Basecode
    , LTRIM(RTRIM(SUBSTRING(Subcode, Number ,CHARINDEX(' ', Subcode + ' ', Number ) - Number))) AS Subcode
FROM @t 
JOIN master..spt_values ON Number <= DATALENGTH(Basecode) + 1  AND type='P'
AND SUBSTRING(' ' + Basecode, Number , 1) = ' ')
,CTE2 AS(
SELECT * 
FROM CTE 
WHERE BaseCode <> @BaseCodetoRemove)

SELECT AccountNumber
    ,STUFF((SELECT  ' ' + CAST(Basecode AS vARCHAR(MAX))
      FROM CTE2 t2 WHERE t1.AccountNumber = t2.AccountNumber
      FOR XML PATH('')),1,1,'')Basecode

    ,STUFF((SELECT ' ' + CAST(Subcode AS vARCHAR(MAX))
      FROM CTE2 t2 WHERE t1.AccountNumber = t2.AccountNumber
      FOR XML PATH('')),1,1,'')Subcode
FROM CTE2  t1
GROUP BY t1.AccountNumber

删除内部以空格分隔的字段部分

1 个答案: