SQL Server : UNION, Columns with Data, Drop the Columns Without - sql-server

Question

I have two tables, LYEAR and CYEAR:
[SKU],[Title],[1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12]
Columns [SKU],[Title],[1],[2],[3],[4],[5],[6],[7],[8] are populated with information from the CYEAR table and [SKU],[Title],[9],[10],[11],[12] from the LYEAR table.
How would I merge the two tables, so that there are columns with NULL values are dropped, leaving me with one row per SKU?
I have got this far:
WITH LYEAR_ORG ([SKU],[Title],[QTY],[DATE]) AS
(
SELECT
T1.ItemNumber, T1.ItemTitle, (SUM(T2.nqty)) AS [UNITS],
CONVERT(VARCHAR(2), (MONTH(T3.dProcessedOn)))
FROM
StockItem T1
LEFT OUTER JOIN
OrderItem T2 ON T1.pkstockItemId = T2.fkStockItemID_processed
LEFT JOIN
[Order] T3 ON T3.pkOrderID = T2.fkOrderID
WHERE
(YEAR(T3.dProcessedOn)) = (YEAR(getdate())-1)
AND (MONTH(T3.dProcessedOn)) > (MONTH(getdate())-1)
GROUP BY
T1.ItemNumber, T1.ItemTitle,
CONVERT(VARCHAR(2), (MONTH(T3.dProcessedOn)))
),
CYEAR_ORG ([SKU],[Title],[QTY],[DATE]) AS
(
SELECT T1.ItemNumber, T1.ItemTitle, (SUM(T2.nqty)) AS [UNITS], CONVERT(varchar(2),(MONTH(T3.dProcessedOn)))
FROM StockItem T1
LEFT OUTER JOIN OrderItem T2 ON T1.pkstockItemId = T2.fkStockItemID_processed
LEFT JOIN [Order] T3 ON T3.pkOrderID = T2.fkOrderID
WHERE (YEAR(T3.dProcessedOn)) = (YEAR(getdate())-1)
AND (MONTH(T3.dProcessedOn)) < (MONTH(getdate()))
GROUP BY T1.ItemNumber, T1.ItemTitle, CONVERT(varchar(2),(MONTH(T3.dProcessedOn)))
),
LYEAR ([SKU],[Title],[1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12]) AS (
SELECT *
FROM LYEAR_ORG
PIVOT ( MAX(QTY) FOR DATE in ([1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12])) as LYEAR_ORDERS
),
CYEAR ([SKU],[Title],[1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12]) AS (
SELECT *
FROM CYEAR_ORG
PIVOT ( MAX(QTY) FOR DATE in ([1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12])) as CYEAR_ORDERS
)
SELECT * FROM LYEAR
UNION ALL
SELECT * FROM CYEAR
I am unable to use INSERT INTO as this is a limitation on the Linnworks platform I am working on, it is SQL Server.

Answer 1

On the assumption that LYEAR columns 1-8 are all NULL, and so are CYEAR 9-12, a JOIN with a ton of COALESCEs, rather than a UNION will get you where you want to be. If, though, there are values for both years in any columns, this will go sideways.
SELECT
l.[SKU]
,l.[Title]
,COALESCE(l.[1], c.[1]) AS [1]
,COALESCE(l.[2], c.[2]) AS [2]
,COALESCE(l.[3], c.[3]) AS [3]
,COALESCE(l.[4], c.[4]) AS [4]
,COALESCE(l.[5], c.[5]) AS [5]
,COALESCE(l.[6], c.[6]) AS [6]
,COALESCE(l.[7], c.[7]) AS [7]
,COALESCE(l.[8], c.[8]) AS [8]
,COALESCE(l.[9], c.[9]) AS [9]
,COALESCE(l.[10], c.[10]) AS [10]
,COALESCE(l.[11], c.[11]) AS [11]
,COALESCE(l.[12], c.[12]) AS [12]
FROM
LYEAR AS l
JOIN
CYEAR AS c
ON
c.SKU = l.SKU
AND c.Title = l.Title;

Answer 2

Here's a solution that will allow for there to be data in both LYEAR and CYEAR, and to switch from one to the other based on the month of the year:
SELECT SKU,
Title,
CASE WHEN DATEPART(MM, GETDATE()) > 1
THEN LYEAR.1
ELSE CYEAR.1
END,
CASE WHEN DATEPART(MM, GETDATE()) > 2
THEN LYEAR.2
ELSE CYEAR.2
END,
CASE WHEN DATEPART(MM, GETDATE()) > 3
THEN LYEAR.3
ELSE CYEAR.3
END,
CASE WHEN DATEPART(MM, GETDATE()) > 4
THEN LYEAR.4
ELSE CYEAR.4
END
--Add rest of CASE statements for remaining months
FROM LYEAR AS L
JOIN CYEAR AS C
ON L.SKU = C.SKU
AND L.Title = C.Title