Question

这是我的问题：

我需要在每个匹配行的数据库中获得第一个结果。我收到了很多行。

原始查询每个custid返回多行：

   select distinct b.custid, name   
        FROM  mth_Charge_Records  a   
        join mth_header b on b.id = a.headerid

所以，为了获得第一行，我想我会给它们编号，然后把它们编号为1.所以这是查询：

  select distinct b.custid, name,   
         count(custid) OVER (order by custid rows unbounded preceding)   as custcount
     FROM  mth_Charge_Records  a   
     join mth_header b on b.id = a.headerid

它有效并且行号被计算在内。所以，我试图将custcount添加为条件，但它失败了：

  select distinct b.custid, name,   
         count(custid) OVER (order by custid rows unbounded preceding)   as custcount
     FROM  mth_Charge_Records  a   
     join mth_header b on b.id = a.headerid     
     where custcount = 1   <<--- Added this line

现在我收到的错误是custcount是一个无效的列。

任何帮助或建议都应该受到赞赏。非常感谢:)。

Answer 1

看看SELECT (Transact-SQL。您会注意到语句的SELECT部分在<{strong> WHERE后评估为。这意味着您无法通过别名引用SELECT中的列。

您也可能注意到您无法在WHERE子句中使用窗口函数。你可以使用CTE，但我想，或许，你真正想要的是一个HAVING条款。因此：

SELECT mt.custid, [name] FROM mth_Charge_Records mCR JOIN mth_header mt on mCR.headerid = mt.id GROUP BY mt.custid, [name] HAVING COUNT(mt.custid) = 1;

编辑：COUNT在重读后有点误导。你想要一个CTE和ROW_NUMBER：

WITH CTE AS( SELECT mt.custid, [name], ROW_NUMBER() OVER (ORDER BY mt.custid) AS RN FROM mth_Charge_Records mCR JOIN mth_header mt on mCR.headerid = mt.id) SELECT * FROM CTE WHERE RN = 1;

Answer 2

您在row_number()的尝试将返回任意行。如果这就是您想要的，为什么不使用聚合？

select h.custid, max(crname)  
from mth_Charge_Records cr join
     mth_header h
     on h.id = cr.headerid;

Answer 3

Try this 
Select a.*
From ( 
select distinct b.custid, name,   
         count(custid) OVER (order by custid rows unbounded preceding)   as custcount
     FROM  mth_Charge_Records  a   
     join mth_header b on b.id = a.headerid 
) a    
     where a.custcount = 1

如何使SQL Over Clause在条件SQL Server中工作

3 个答案: