如何在一列中总计行组的结果？

Question

我的表看起来像：

Customer   Error1   Error2   Error3   Error4  TotalErrors  TotalOrders 
CustomerName [SUM(Error1)] [SUM(Error2)] [SUM(Error3)] [SUM(Error3) [SUM(TotalErrors)] [TotalOrders] 
Total 

的中间行客户名称是行集团的一部分下令CustomerWithErrors简单地告知用户，如果给定的客户名称有任何的4个错误的。如果没有，它不会显示在表格中。

在最后一行中，我只需右键单击并为行组中的每列添加总计。这完全按照我的意愿输出每列的总和，除了最后一列。当我添加总数时，它给了我所有客户名称的TotalErrors字段的总和，而不仅仅是那些有错误的字段（表中实际显示的字段）。因此，这个数字非常庞大。我怎样才能得到总错误列中显示的数字总和？

编辑：下面是设计和预览模式一些截图：

设计：http://i.imgur.com/6kEmzw3.png

前瞻：http://i.imgur.com/NOpWmAI.png

编辑2：SQL的顶级表矩阵：

一些这对我来说是新的，所以它可能看起来不太好。让我知道我能解释什么。谢谢。

SELECT 
     PKEY 
     ,Customer 
     ,isnull(CustomerName, 'Unknown') AS CustomerName 
     ,CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,CASE WHEN ErrorCategory='Success' 
     THEN 0 
     ELSE  
     COUNT(NotLate) 
     END AS OnTimeCount 
     ,COUNT(Invalid) AS InvalidCount 
     ,COUNT(WrongData) AS WrongDataCount 
     ,COUNT(Duplicate) AS DuplicateCount 
     ,COUNT(MissingData) AS MissingDataCount 
     ,(COUNT(Invalid)+COUNT(WrongData)+COUNT(Duplicate)+COUNT(MissingData)) AS TotalErrors 
     ,Total945s AS TotalReport945s 
     --,COUNT(*) OVER(PARTITION BY [CustomerWithErrors]) Total945s 
     , (SELECT (Total945s) WHERE CustomerWithErrors>=0) AS TotalErrors 

     FROM(

    SELECT 
     PKEY 
     ,Customer 
     ,CASE WHEN CustomerName = '' OR CustomerName IS NULL 
     THEN 'Unknown' 
     ELSE CustomerName 
     END AS CustomerName 
     ,CASE WHEN CustomerName='' OR CustomerName IS NULL OR CustomerName='Unknown' 
     THEN 0 
     ELSE 
     (SELECT DISTINCT Customer WHERE ErrorCategory <> 'Success') END 
     AS CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,CASE WHEN ErrorCategory='Success' 
     THEN 'N' 
     ELSE OnTime   
     END AS OnTime 
     ,(SELECT Customer WHERE OnTime = 'Y') AS NotLate 
     ,(SELECT Customer WHERE ErrorCategory = 'Invalid') AS Invalid 
     ,(SELECT Customer WHERE ErrorCategory = 'Wrong Data') AS WrongData 
     ,(SELECT Customer WHERE ErrorCategory = 'Duplicate') AS Duplicate 
     ,(SELECT Customer WHERE ErrorCategory = 'Missing Data') AS MissingData 
     , COUNT(*) OVER(PARTITION BY [Customer]) Total945s 



    FROM EDI945UTIErrors 


    ) S 


     GROUP BY 
     PKEY 
     ,Customer 
     ,CustomerName 
     ,CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,OnTime 
     ,Total945s 

编辑3： 这是否看起来像我应该在报告中聚集做什么：与PIVOT SQL？谢谢。

SELECT 
     S.PKEY 
     ,Customer 
     ,isnull(S.CustomerName, 'Unknown') AS CustomerName 
     ,CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,CASE WHEN ErrorCategory='Success' 
     THEN 0 
     ELSE  
     COUNT(NotLate) 
     END AS OnTimeCount 
     ,Invalid 
     ,WrongData 
     ,Duplicate 
     ,MissingData 
     --,COUNT(Invalid) AS InvalidCount 
     --,COUNT(WrongData) AS WrongDataCount 
     --,COUNT(Duplicate) AS DuplicateCount 
     --,COUNT(MissingData) AS MissingDataCount 
     --,(COUNT(Invalid)+COUNT(WrongData)+COUNT(Duplicate)+COUNT(MissingData)) AS TotalErrors 
     ,Total945s AS TotalReport945s 

     --,COUNT(*) OVER(PARTITION BY [CustomerWithErrors]) Total945s 
     --, (SELECT (Total945s) WHERE CustomerWithErrors>=0) AS TotalErrors 

     FROM(

    SELECT 
     PKEY 
     ,Customer 
     ,CASE WHEN CustomerName = '' OR CustomerName IS NULL 
     THEN 'Unknown' 
     ELSE CustomerName 
     END AS CustomerName 
     ,CASE WHEN CustomerName='' OR CustomerName IS NULL OR CustomerName='Unknown' 
     THEN 0 
     ELSE 
     (SELECT DISTINCT Customer WHERE ErrorCategory <> 'Success') END 
     AS CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,CASE WHEN ErrorCategory='Success' 
     THEN 'N' 
     ELSE OnTime   
     END AS OnTime 
     ,(SELECT Customer WHERE OnTime = 'Y') AS NotLate 
     --,(SELECT Customer WHERE ErrorCategory = 'Invalid') AS Invalid 
     --,(SELECT Customer WHERE ErrorCategory = 'Wrong Data') AS WrongData 
     --,(SELECT Customer WHERE ErrorCategory = 'Duplicate') AS Duplicate 
     --,(SELECT Customer WHERE ErrorCategory = 'Missing Data') AS MissingData 
     , COUNT(*) OVER(PARTITION BY [Customer]) Total945s 



    FROM EDI945UTIErrors 
    --WHERE Customer ='104646' and OnTime = 'Y' 
    --WHERE CUSTOMER='107834' 
    --AND ErrorCategory = 'Wrong Data' 
    --AND ERRORMESSAGE LIKE '%Overshipment Not Allowed%' 

    ) S 
    LEFT JOIN 
    (
    SELECT PKEY, ISNULL(CustomerName, 'Unknown') AS CustomerName, [Invalid], [WrongData], [Duplicate], [MissingData], [Success] 
FROM EDI945UTIErrors 
PIVOT (Count(Customer) FOR ErrorCategory in ([Invalid], [WrongData], [Duplicate], [MissingData], [Success])) 
AS P) AS PT ON S.PKEY = PT.PKEY 

     GROUP BY 
     S.PKEY 
     ,Customer 
     ,S.CustomerName 
     ,CustomerWithErrors 
     ,Warehouse 
     ,ErrorMessage 
     ,PROCESSDATE 
     ,CARTONERRORMESSAGE 
     ,Division 
     ,BOLNO 
     ,BOLSHIPDATE 
     ,CARTONDETAILERRMSG 
     ,ErrorCategory 
     ,OnTime 
     ,Total945s 
     ,Invalid 
     ,WrongData 
     ,MissingData 
     ,Duplicate 

非常感谢您的宝贵时间。

Answer 1

从发布的图片看，您的数据集可能会吸引所有客户的订单，而不仅仅是那些有错误的订单。检查这一点的一个好方法是从数据集中获取SQL并在Management Studio中运行它。如果您获取的是不想计数的行，则可以在Tablix上使用过滤器。

要在Tablix中应用过滤器，右击它并转到Tablix属性：

点击过滤器，你可以选择像只具有一个错误标志命令：

如果这不能解决您的问题，请让我知道并将您的数据集的SQL添加到您的问题中，然后我们将返回到绘图板。 :)

编辑 - 11/12/2014

感谢您的SQL！如果CustomerWithErrors仅返回客户编号，那么过滤器将无济于事。你仍然在拉每个记录（甚至那些ErrorCategory ='成功'）。你知道你有两列别名为TotalErrors吗？有

(COUNT(Invalid)+COUNT(WrongData)+COUNT(Duplicate)+COUNT(MissingData)) AS TotalErrors 

和

(SELECT (Total945s) WHERE CustomerWithErrors>=0) AS TotalErrors 

第一个看起来像它可能工作，但我不知道你正在使用你的表，。另一个需要思考的问题是，你正在计算聚合总量。你可能最终相乘他们，如果你遇到，你有这样的结果集中的情况：

Customer | ErrorCategory | TotalErrors 
1234  | Invalid   | 2 
1234  | Duplicate  | 2 

结果，你会在您的报告中看到：

Customer | TotalErrors 
1234  | 4 

如果你想避免，你可以让SSRS进行汇总。你有没有考虑过做一个PIVOT？ 你可以做这样的事情：

Select Customer, ISNULL(CustomerName, 'Unknown'), [Invalid], [Wrong Data], [Duplicate], [Missing Data], [Success] 
FROM EDITable 
PIVOT (Count(Customer) FOR ErrorCategory in ([Invalid], [Wrong Data], [Duplicate], [Missing Data], [Success])) 

然后在你的总错误报告，使用显示总和（无效）+总和（错误的数据）+等的表达。只要每个客户有一个记录，您的数据就不会翻倍，您仍然可以获得非错误记录的总数。

我希望帮助。让我知道你是否有疑问！