避免重复行到 SUM 两次

Question

我有一个包含重复行的表，我如何将该重复行求和为不同的值？我知道我可以将不同的数据插入新表并以这种方式求和，但我试图避免创建任何类型的表。 表：

SELECT '123' DateID,'8' HOUR,'6000' AMT, '20150801' PayEndingDate, '20150807' CheckDate
INTO #DATA
UNION ALL
SELECT '456',8,'6000','20150801','20150807' UNION ALL
SELECT '789',8,'6000','20150801','20150807' UNION ALL --DUP ROW
SELECT '789',8,'6000','20150801','20150807' UNION ALL --DUP ROW
SELECT '987',8,'6000','20150725','20150807' UNION ALL
SELECT '654',8,'6000','20150725','20150807' UNION ALL
SELECT '321',8,'6000','20150725','20150807' UNION ALL --DUP ROW
SELECT '321',8,'6000','20150725','20150807' --DUP ROW
SELECT * FROM #DATA

|DateID | HOUR  | AMT      | PayEndingDate  | CheckDate
 123      8       6000       20150801         20150807
 456      8       6000       20150801         20150807
 789      8       6000       20150801         20150807
 789      8       6000       20150801         20150807
 987      8       6000       20150725         20150807
 654      8       6000       20150725         20150807
 321      8       6000       20150725         20150807
 321      8       6000       20150725         20150807

SELECT  DISTINCT SUM(HOUR) AS HOURS, AMT, PayEndingDate, CHECKDATE
FROM #DATA
GROUP BY AMT,PayEndingDate,CHECKDATE

| HOUR  | AMT      | PayEndingDate  | CheckDate
  32      6000       20150725         20150807
  32      6000       20150801         20150807

如您所见，ID 321 和 789 重复，导致小时数为 32 而应为 24。在 SUM 中添加 distinct 仅返回 8。在这种情况下如何避免重复？

Answer 1

试试这个...

Select SUM(HOUR) AS HOURS, AMT, PayEndingDate, CHECKDATE
From (
SELECT  DISTINCT DateID, HOUR , AMT, PayEndingDate, CHECKDATE
FROM #DATA
) A
GROUP BY AMT,PayEndingDate,CHECKDATE

Answer 2

DISTINCT 本身可能有点贵。就个人而言，我更喜欢使用 A CTE 和 ROW_NUMBER：

WITH CTE AS(
    SELECT *,
           ROW_NUMBER() OVER (PARTITION BY DateID ORDER BY DateID) AS RN
    FROM #Data)
SELECT SUM(HOUR) AS HOURS, AMT, PayEndingDate, CHECKDATE
FROM CTE
WHERE RN = 1
GROUP BY AMT, PayEndingDate, CHECKDATE;