quote:

---

Originally posted by sqlclarify

I know I asked this question before but I can't find it.. and I have this problem again.. Lets say I have

SELECT
COLA,
COLB,
COLC,
COLQ,
FROM xyz....

I need to group by only COLA AND COL B. But according to the syntax I have to list all the columns. But if I do that it changes the meaning of the grouping. How do I avoid that? How do I get by this problem?

Thank you.

---

1.SELECT COLA,COLB,....
FROM
(
SELECT ROW_NUMBER() OVER (PARTITION BY COLA,COL B ORDER BY IDCol) AS Seq,
COLA,
COLB,
COLC,
COLQ,
FROM xyz....
)t
WHERE Seq=1

2.SELECT COLA,COLB,....
FROM
(
SELECT ROW_NUMBER() OVER (PARTITION BY COLA,COL B ORDER BY IDCol DESC) AS Seq,
COLA,
COLB,
COLC,
COLQ,
FROM xyz....
)t
WHERE Seq=1

3.SELECT COLA,COLB,....
FROM
(
SELECT ROW_NUMBER() OVER (PARTITION BY COLA,COL B ORDER BY NEWID()) AS Seq,
COLA,
COLB,
COLC,
COLQ,
FROM xyz....
)t
WHERE Seq=1