试图找出如何做一个dynamic的数据透视表连接
我还是一个T-SQL的初学者，有一些相对简单的select语句和一些连接。 我想把它放到一个dynamic数据透视表中。 这个声明的目的是要显示非营利性服务及其结果，新旧结果是相同数据的不同词语的组合。
SELECT Service.Name, OutcomeIndicator."Desc" AS OutcomeDesc, NewOutcomeIndicator.Description As NewOutcomeDesc FROM service FULL OUTER JOIN OutcomeIndicator ON ServiceOutcomeIndicator.OutcomeIndicatorID=OutcomeIndicator.OutcomeIndicatorID FULL OUTER JOIN NewOutcomeIndicator ON NewOutcomeIndicatorMapping.NewOutcomeIndicatorID=NewOutcomeIndicator.NewOutcomeIndicatorID 但是，当我运行这个语句时，显然会返回每一行，包括ServiceName的多个值，因为每个服务都可以有多个结果/新对象配对。
+-------------+----------------+------------------+ | ServiceName | OutcomeDesc1 | NewOutcomeDesc1 | +-------------+----------------+------------------+ | Service34 | OutcomeDesc14 | NewOutcomeDesc14 | | Service34 | OutcomeDesc12 | NewOutcomeDesc12 | | Service35 | OutcomeDesc4 | NewOutcomeDesc4 | | Service44 | OutcomeDesc80 | NewOutcomeDesc80 | | Service44 | OutcomeDesc99 | NewOutcomeDesc99 | | Service51 | OutcomeDesc50 | NewOutcomeDesc50 |
出于报告目的，因为这是复制到Excel中，我想要更多的东西是这样的：
+--------------+-----------------+-------------------+-----------------+-------------------+--------------+---------------------+ | ServiceName | OutcomeDesc1 | NewOutcomeDesc1 | OutcomeDesc2 | NewOutcomeDesc2 | OutcomeDesc3 | NewOutcomeDesc3 | +--------------+-----------------+-------------------+-----------------+-------------------+--------------+---------------------+ | Service34 | OutcomeDesc14 | NewOutcomeDesc14 | OutcomeDesc12 | NewOutcomeDesc12 | | | | Service35 | OutcomeDesc4 | NewOutcomeDesc4 | | | | | | Service44 | OutcomeDesc80 | NewOutcomeDesc80 | OutcomeDesc99 | NewOutcomeDesc99 | | | | Service51 | OutcomeDesc50 | NewOutcomeDesc50 | | | | | +--------------+-----------------+-------------------+-----------------+-------------------+--------------+---------------------+
任何想法如何可以更改我的SQL语句来做到这一点？ 我可以静态定义列的数量，因为我知道每个服务的结果不会超过3个。
我知道如何在sql中做一个静态数据透视表，但是因为有数百个服务和结果，所以我无法定义统计分配每个Service值到一个列的逻辑。 dynamic的数据透视表有点难以理解，但我还没有find一个教程来做这个，同时还包括一个join语句。
任何帮助赞赏。
你不需要dynamic的pivot，实际上你甚至不需要使用真正的pivot命令，你可以很容易地用CASE和GROUP BY来模拟
在检索结果之后，使用ROW_NUMBER（）函数为每行分配数字1-3，最后使用该RNselect进入哪个列的内容。
这是一个片段，在第一个CTE中添加真实的代码。
WITH CTE_Source AS ( --Your query goes here ) , CTE_RN AS ( SELECT * , ROW_NUMBER() OVER (PARTITION BY ServiceName ORDER BY OutcomeDesc) RN FROM CTE_Source ) SELECT ServiceName , MAX(CASE WHEN RN = 1 THEN OutcomeDesc END) AS OutcomeDesc1 , MAX(CASE WHEN RN = 1 THEN NewOutcomeDesc END) AS NewOutcomeDesc1 , MAX(CASE WHEN RN = 2 THEN OutcomeDesc END) AS OutcomeDesc2 , MAX(CASE WHEN RN = 2 THEN NewOutcomeDesc END) AS NewOutcomeDesc2 , MAX(CASE WHEN RN = 3 THEN OutcomeDesc END) AS OutcomeDesc3 , MAX(CASE WHEN RN = 3 THEN NewOutcomeDesc END) AS NewOutcomeDesc3 FROM CTE_RN GROUP BY ServiceName
使用您的数据作为SourceTable的DEMO