我可以將兩列合併爲一列，然後將它們的值計算爲一列嗎？

Question

我目前使用這個SQL語句返回由誰完成每個記錄計數（RFI）：

SELECT Completed_By, COUNT(DISTINCT Unique_ID) AS RFICount FROM RFI_ 
WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
GROUP BY Completed_By 

這將返回一個數據集，像下面這樣：

Completed_By  RFICount 
------------  --------- 
SMITH, Bob  1 
DOE, John  15 
WARNER, Jack  11 

這是我多麼希望它看起來。但是，有人可以是完成或協助完成記錄（RFI）的主要或次要人員。無論他們是完成記錄（Completed_By）還是輔助它（Assisting_Analyst），我都想在一列中全部顯示。

當我使用下面的SQL語句，它顯示了這些作爲單獨的，當他們需要進行彙總：

SELECT Completed_By, Assisting_Analyst, COUNT(DISTINCT Unique_ID) AS RFICount 
FROM RFI_ WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
GROUP BY Completed_By, Assisting_Analyst 

這將返回以下類型的輸出：

Completed_By  Assisting_Analyst  RFICount 
------------  -----------------  -------- 
SMITH, Bob  NULL     1 
DOE, John  NULL     15 
WARNER, Jack  NULL     7 
WARNER, Jack  SMITH, Bob   4 

因爲鮑勃·史密斯協助傑克華納有四個記錄（RFIs），有一個獨立的行創建。但我只是希望每個拿到個人信用，並讓它在一列中顯示出來，就像這樣：

Completed_By  RFICount 
------------  --------- 
SMITH, Bob  5 
DOE, John  15 
WARNER, Jack  11 

在上面的例子中，四個記錄他們的工作就共同記入他們兩個。當然，這意味着一些記錄被計數兩次，但這是我想要的這個應用程序。

我嘗試過工會和一堆其他的東西，但它要麼不給我我想要的東西，因爲我想顯示或它給了我一個沒有數據的錯誤。我知道DISTINCT必須去，但不知道如何完全做到這一點。這是在SQL Server 2008 R2中完成的。幫幫我？

Answer 1

SELECT t.Completed_By, COUNT(DISTINCT t.Unique_ID) AS RFICount 
    FROM (SELECT Completed_By, Unique_ID 
       FROM RFI_ 
       WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
      UNION ALL 
      SELECT Assisting_Analyst AS Completed_By, Unique_ID 
       FROM RFI_ 
       WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
        AND Assisting_Analyst IS NOT NULL) t 
    GROUP BY t.Completed_By 

Answer 2

這樣做的最好方法可能是彙總，避免了子查詢，並可能在大型數據集上執行得更好。

SELECT Completed_By, Assisting_Analyst, COUNT(DISTINCT Unique_ID) AS RFICount 
FROM RFI_ WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
AND Assisting_Analyst IS NULL and Completed_By IS NOT NULL 
GROUP BY Completed_By, Assisting_Analyst WITH ROLLUP 

ROLLUP意味着總值應該包含在Completed_By行中，因此排除了where子句中的輔助行。

Answer 3

SELECT a.Completed_By 
    , COUNT(DISTINCT a.Unique_ID) 
     + (SELECT COUNT(DISTINCT b.Unique_ID) 
      FROM RFI_ b 
      WHERE b.Date_Submitted BETWEEN '20110101' AND '20110630' 
      AND b.Assisting_Analyst = a.Completed_By 
     ) 
     AS RFICount 
FROM RFI_ a 
WHERE a.Date_Submitted BETWEEN '20110101' AND '20110630' 
GROUP BY a.Completed_By 

闖闖：

SELECT COALESCE(first.Completed_By, second.Assisting_Analyst) 
     AS Completed_By 
    , COALESCE(first.RFICount,0) + COALESCE(second.RFICount,0) 
     AS RFICount 
FROM 
    (SELECT Completed_By 
     , COUNT(DISTINCT Unique_ID) AS RFICount 
    FROM RFI_ 
    WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
    GROUP BY Completed_By 
) AS first 
FULL JOIN 
    (SELECT Assisting_Analyst 
     , COUNT(DISTINCT Unique_ID) AS RFICount 
    FROM RFI_ 
    WHERE Date_Submitted BETWEEN '20110101' AND '20110630' 
    GROUP BY Assisting_Analyst 
) AS second 
ON first.Completed_By = second.Assisting_Analyst