排除结果中的NULL列

Question

我的查询会返回这些结果

SELECT 
     BorrowerNumber
     ,CASE WHEN [Score_CorrectID_Accuracy] = 'Error' THEN 'Error' END AS [Correct ID]
     ,CASE WHEN [Score_ProperlyIdentified_Accuracy] = 'Error' THEN 'Error' END AS [PCA Identified Itself]
     ,CASE [Score_Tone_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Tone_Accuracy]
     ,CASE [Score_Accuracy_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Accuracy_Accuracy]
     ,CASE [Score_Notepad_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Notepad_Accuracy] 
    FROM 
        Reviews INNER JOIN QCTier1 ON Reviews.ReviewID = QCTier1.ReviewID

这些是我希望实现的结果，其中最后一列不在结果中，因为它具有NULL值。感谢

Answer 1

据我所知，标准的tsql无法根据值显示/隐藏列。

如果您正在编程，可以在其他程序/ UI中实现此功能。

还有另一种有趣的方法可以解决这个问题，使用iff并在stackoverflow中进行透视，但这对于这种业务需求来说太复杂了！

也许您应该尝试在报表程序中实现此功能，而不是在SQL中。

Answer 2

试试这个：

DECLARE @COUNT int

SELECT @COUNT = COUNT(*)
FROM Reviews INNER JOIN QCTier1 ON Reviews.ReviewID = QCTier1.ReviewID
WHERE Score_Notepad_Accuracy IS NOT NULL

IF @COUNT = 0 THEN
    SELECT 
       BorrowerNumber
      ,CASE WHEN [Score_CorrectID_Accuracy] = 'Error' THEN 'Error' END AS [Correct ID]
      ,CASE WHEN [Score_ProperlyIdentified_Accuracy] = 'Error' THEN 'Error' END AS [PCA Identified Itself]
      ,CASE [Score_Tone_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Tone_Accuracy]
      ,CASE [Score_Accuracy_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Accuracy_Accuracy]
    FROM 
       Reviews INNER JOIN QCTier1 ON Reviews.ReviewID = QCTier1.ReviewID
ELSE
    SELECT 
       BorrowerNumber
      ,CASE WHEN [Score_CorrectID_Accuracy] = 'Error' THEN 'Error' END AS [Correct ID]
      ,CASE WHEN [Score_ProperlyIdentified_Accuracy] = 'Error' THEN 'Error' END AS [PCA Identified Itself]
      ,CASE [Score_Tone_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Tone_Accuracy]
      ,CASE [Score_Accuracy_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Accuracy_Accuracy]
    FROM 
       Reviews INNER JOIN QCTier1 ON Reviews.ReviewID = QCTier1.ReviewID

我个人不会这样做。您应该在报告程序中隐藏该列，而不是在数据源中隐藏。

Answer 3

如果不使用一些非常古怪和困难的代码，您将无法从结果集中排除列本身。但是你可以做的（几乎每个报告工具）都是根据公式甚至单个值显示或隐藏报表中的列。

因此，请使用现有查询，对于您可能要隐藏的每个列，添加另一列，将“隐藏列”聚合在整个数据集上。

SELECT sq.*, max(sq.[Correct ID]) as MAX_CORRECT_ID, max(sq.[PCA Identified Itself]) as MAX_PCA_ID_SELF -- ... rinse and repeat FROM ( SELECT BorrowerNumber ,CASE WHEN [Score_CorrectID_Accuracy] = 'Error' THEN 'Error' END AS [Correct ID] ,CASE WHEN [Score_ProperlyIdentified_Accuracy] = 'Error' THEN 'Error' END AS [PCA Identified Itself] ,CASE [Score_Tone_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Tone_Accuracy] ,CASE [Score_Accuracy_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Accuracy_Accuracy] ,CASE [Score_Notepad_Accuracy] WHEN 'Error' THEN 'Error' END AS [Score_Notepad_Accuracy] FROM Reviews INNER JOIN QCTier1 ON Reviews.ReviewID = QCTier1.ReviewID ) sq

任何时候MAX列在第一行中为空，您可以确定相应列中没有非空值。

在我看来，对于通常只返回几行的结果集，这是一个巨大的时间浪费，因为您可以在报表工具中使用简单的聚合或公式来有条件地隐藏列，并保持SQL清洁。如果数据集中有数十万行，或者您有更复杂的需求需要分组或窗口，或两者兼而有之，那么这可能是值得的。例如，有时显示与当前行共享特定属性的所有行的聚合数据，而不按该列对整个报表进行分组。