How to ADD latest values if not 0 [closed] - sql-server

Question

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 3 years ago.
Improve this question
Is there a better way of adding the latest 3 values in a row where the value is not 0 (i.e, skip column when value = 0)?
Example :
[Jan] [Feb] [Mar] [Apr] [May]
[2] [3] [10] [0] [7]
[2] [3] [10] [0] [0]
[2] [3] [0] [0] [7]
[2] [3] [10] [0] [7]
Looking for the row values; 20, 15, 12, 20.
At the moment I'm only thinking of a massive case statement...

Answer 1

You can do this by unpivoting the columns ordering them and then summing like this:
SELECT *
FROM MyRows AS M
CROSS APPLY
(
SELECT SUM(ColumnValue) AS SumValue
FROM
(
SELECT TOP 3 V.ColumnValue
FROM (
VALUES
(1,Column5)
,(2,Column4)
,(3,Column3)
,(4,Column2)
,(5,Column1)
) AS V(RankMe,ColumnValue)
WHERE V.ColumnValue <> 0
ORDER BY V.RankMe
) AS SumValues
) AS TopThree
Example at dbfiddle.uk
However as others point out we're unsure why you'd want to do this, I suspect your data model has issues and you'd be better with these values in columns to start with.

Answer 2

Assuming there is an ID column, and assuming the columns go up to Dec.
You need to unpivot your data and normalise it. Then you can easily aggregate it. This is how you would do it "on the fly":
WITH CTE AS(
SELECT YT.ID,
V.MonthNo,
V.[MonthName],
V.[Value],
ROW_NUMBER() OVER (PARTITION BY YT.ID, IIF(V.[Value] = 0,0,1) ORDER BY V.MonthNo DESC) AS RN
FROM dbo.YourTable YT
CROSS APPLY (VALUES('Jan',1,Jan),
('Feb',2,Feb),
('Mar',3,Mar),
('Apr',4,Apr),
('May',5,May),
('Jun',6,Jun),
('Jul',7,Jul),
('Aug',8,Aug),
('Sep',9,Sep),
('Oct',10,Oct),
('Nov',11,Nov),
('Dec',12,[Dec]))V([MonthName], MonthNo, [Value]))
SELECT YT.ID,
SUM(V.[Value]) AS [Aggregate]
FROM CTE
WHERE RN <= 3
GROUP BY YT.ID;
Ideally, though, you should be fixing your data model.

Answer 3

You could convert columns to rows using CROSS APPLY and aggregate:
WITH cte AS (
SELECT id, ROW_NUMBER() OVER (PARTITION BY id ORDER BY mnum DESC) AS rn, mval
FROM t
CROSS APPLY (VALUES
(1 , Jan),
(2 , Feb),
(3 , Mar),
(4 , Apr),
(5 , May),
(6 , Jun),
(7 , Jul),
(8 , Aug),
(9 , Sep),
(10, Oct),
(11, Nov),
(12, Dec)
) x(mnum, mval)
WHERE mval <> 0
)
SELECT id, SUM(mval)
FROM cte
WHERE rn <= 3
GROUP BY id

Answer 4

Try this:
DECLARE #DataSource TABLE
(
[RowID] INT IDENTITY(1,1)
,[Jan] INT
,[Feb] INT
,[Mar] INT
,[Apr] INT
,[May] INT
);
INSERT INTO #DataSource ([Jan], [Feb], [Mar], [Apr], [May])
VALUES (2, 3, 10, 0, 7)
,(2, 3, 10, 0, 0)
,(2, 3, 0, 0, 7)
,(2, 3, 10, 0, 7);
WITH DataSource AS
(
SELECT [RowID]
,ROW_NUMBER() OVER (PARTITION BY [RowID] ORDER BY CASE [column]
WHEN 'Jan' THEN 1
WHEN 'Feb' THEN 2
WHEN 'Mar' THEN 3
WHEN 'Apr' THEN 4
WHEN 'May' THEN 5
END DESC
) AS [MonthID]
,[value]
FROM #DataSource
UNPIVOT
(
[value] FOR [column] IN ([Jan], [Feb], [Mar], [Apr], [May])
) UNPVT
WHERE [value] <> 0
)
SELECT [RowID]
,SUM([value])
FROM DataSource
WHERE [MonthID] <= 3
GROUP BY [RowID];
The idea is simple:
unpivot the data
order the months from latest to newest excluding months without values
sum the values but for the last 3 months only