Question

I have a query that returns a result set similar to the one below (in reality it is far bigger, thousands of rows):

      A    | B  | C  |    D      -----|----|----|----- 1    NULL | d0 | d0 | NULL 2    NULL | d0 | d1 | NULL 3    NULL | d0 | d2 |   a0 4      d0 | d1 | d1 | NULL 5      d0 | d2 | d2 |   a0 

Two of the rows are considered duplicates, 1 and 2, because A, B and D are the same. To eliminate this, I could use SELECT DISTINCT A, B, D but then I do not get column C in my result set. Column C is necessary information for rows 3, 4 and 5.

So how do I come from the result set above to this one (the result appearing in C4 can also be NULL instead of d1):

      A    | B  | C    | D      -----|----|------|----- 1    NULL | d0 | NULL | NULL 3    NULL | d0 | d2   |   a0 4      d0 | d1 | d1   | NULL 5      d0 | d2 | d2   |   a0 

Answer 1

DECLARE @YourTable TABLE (   A VARCHAR(2)   , B VARCHAR(2)   , C VARCHAR(2)   , D VARCHAR(2))  INSERT INTO @YourTable VALUES (NULL, 'd0', 'd0', NULL) INSERT INTO @YourTable VALUES (NULL, 'd0', 'd1', NULL) INSERT INTO @YourTable VALUES (NULL, 'd0', 'd2', 'a0') INSERT INTO @YourTable VALUES ('d0', 'd1', 'd1', NULL) INSERT INTO @YourTable VALUES ('d0', 'd2', 'd2', 'a0')   SELECT A, B, C = MIN(C), D FROM @YourTable GROUP BY A, B, D 

---

SELECT A, B, CASE WHEN MIN(C) = MAX(C) THEN MIN(C) ELSE NULL END, D FROM @YourTable GROUP BY A, B, D 

---

SELECT A, B, CASE WHEN MIN(COALESCE(C, 'dx')) = MAX(COALESCE(C, 'dx')) THEN MIN(C) ELSE NULL END, D FROM @YourTable GROUP BY A, B, D