Reputation: 185
I'm given a messy "fact" table in SQL Server 2012 with OrderId and ProductId being the keys. The same OrderId can have multiple ProductId (up to 5 rows). For reporting purposes I need to leave only one row for each combination OrderId+ProductId, using the max value from the 5 fields Ask, Buy, Rec, Pen, Ret.
I've tried using the pivot operator over ProductId, but don't know how to get the max value from the 5 different fields.
Sample data:
DROP TABLE #tmpSO;
CREATE TABLE #tmpSO
(OrderId INT, ProductId INT, Ask INT, Buy INT, Rec INT, Pen INT, Ret INT);
INSERT INTO #tmpSO (OrderId, ProductId, Ask, Buy, Rec, Pen, Ret)
VALUES
(67,1,2,0,0,2,0),
(67,4,30,0,0,30,0),
(67,4,0,30,0,30,0),
(67,6,20,0,0,0,0),
(67,6,0,20,0,0,0),
(67,6,0,0,20,0,0),
(67,9,30,0,0,0,0),
(67,9,0,30,0,0,0),
(67,9,0,0,25,0,0),
(67,9,0,0,0,5,0);
SELECT * FROM #tmpSO;
OrderId ProductId Ask Buy Rec Pen Ret
67 1 2 0 0 2 0
67 4 30 0 0 30 0
67 4 0 30 0 30 0
67 6 20 0 0 0 0
67 6 0 20 0 0 0
67 6 0 0 20 0 0
67 9 30 0 0 0 0
67 9 0 30 0 0 0
67 9 0 0 25 0 0
67 9 0 0 0 5 0
My desired result would be:
OrderId ProductId Ask Buy Rec Pen Ret
67 1 2 0 0 2 0
67 4 30 30 0 30 0
67 6 20 20 20 0 0
67 9 30 30 25 5 0
Please advise if this is feasible using Pivot (or other operator), and a help showing how to use it. Many thanks.
Upvotes: 1
Views: 56
Reputation: 4039
Why can't you use GROUP BY
and MAX()
?
It would give your expected result:
select orderid, productid, MAX(ask) ask, MAX(buy) as buy, MAX(rec) as rec, MAX(pen) as pen, MAX(ret) as ret
from #tmpso
group by orderid, productid
order by orderid, productid
Working copy here.
Upvotes: 2