MS SQL Combining 5 columns to 3 columns - sql

Question

I am trying to figure out how to combine 5 columns to 3 columns and avoiding duplicities at the same time.
I have been thinking about it for a few days and I can't figure out any solution so far, so I have decided to ask for help here.
Basically right now we use 5 columns in our database let's say A,B,C,D,E and somebody from our company has decided that we will be moving this information to new 3 columns let's call them NEW_1,NEW_2,NEW_3, because there are no cases when we would have more than 3 values per row in the DB.
I need to somehow extract those 5 columns and get them to just 3 columns without repeating them.
So far I was thinking about and trying using CASE, but I can't figure it out how to stop it from showing the same value in all 3 NEW columns. I was thinking about assigning some variables to determine which columns to skip in select CASE, but I have found out that I can't do it like that.
If someone could at least direct me to the right way I would really appreciate it.
SELECT CASE
WHEN A is not null THEN A
WHEN B is not null THEN B
WHEN C is not null THEN C
WHEN D is not null THEN D
WHEN E is not null THEN E
ELSE NULL
END as NEW_1,
CASE
WHEN B is not null THEN B
WHEN C is not null THEN C
WHEN D is not null THEN D
WHEN E is not null THEN E
ELSE NULL
END as NEW_2,
CASE
WHEN C is not null THEN C
WHEN D is not null THEN D
WHEN E is not null THEN E
ELSE NULL
END as NEW_3

Answer 1

Here is a option where we unpivot the data and then apply a conditional aggregation within a CROSS APPLY
Example
Declare #YourTable Table ([A] varchar(50),[B] varchar(50),[C] varchar(50),[D] varchar(50),[E] varchar(50))
Insert Into #YourTable Values
(1,2,3,NULL,NULL)
,(NULL,4,NULL,5,NULL)
,(NULL,null,6,7,8)
,(9,NULL,NULL,NULL,NULL)
Select A.*
,B.*
From #YourTable A
Cross Apply (
Select Val1 = max(case when ColNr=1 then Value end)
,Val2 = max(case when ColNr=2 then Value end)
,Val3 = max(case when ColNr=3 then Value end)
From (
Select ColNr = row_number() over (order by Seq)
,B1.*
From (values (1,A)
,(2,B)
,(3,C)
,(4,D)
,(5,E)
) B1(Seq,Value)
Where Value is not null
) B2
) B
Returns
Just for Fun, here is an XML version
Select A.*
,Val1 = XMLData.value('/x[1]','varchar(max)')
,Val2 = XMLData.value('/x[2]','varchar(max)')
,Val3 = XMLData.value('/x[3]','varchar(max)')
From #YourTable A
Cross Apply ( values ( convert(xml,
concat('<x>'+A+'</x>'
,'<x>'+B+'</x>'
,'<x>'+C+'</x>'
,'<x>'+D+'</x>'
,'<x>'+E+'</x>'
) ) ) ) B(XMLData)

Answer 2

I like John's answer and probably better than mine, but here's a slightly different version if the columns are fixed to a specific number.
Build a delimited string of the column values, then use xml to extract the 1st, 2nd, 3rd values.
Declare #YourTable Table ([A] varchar(50),[B] varchar(50),[C] varchar(50),[D] varchar(50),[E] varchar(50));
Insert Into #YourTable Values
(1,2,3,NULL,NULL),(NULL,4,NULL,5,NULL),(NULL,null,6,7,8),(9,NULL,NULL,NULL,NULL);
WITH CTE AS (
SELECT CASE WHEN a IS NULL THEN '' ELSE a + '~' END +
CASE WHEN b IS NULL THEN '' ELSE b + '~' END +
CASE WHEN c IS NULL THEN '' ELSE c + '~' END +
CASE WHEN d IS NULL THEN '' ELSE d + '~' END +
CASE WHEN e IS NULL THEN '' ELSE e + '~' END AS Combined
FROM #YourTable
)
SELECT
ISNULL(CAST(N'<x>' + REPLACE(Combined, '~', N'</x><x>') + N'</x>' AS XML).value('/x[1]', 'nvarchar(max)'), '') [new_1],
ISNULL(CAST(N'<x>' + REPLACE(Combined, '~', N'</x><x>') + N'</x>' AS XML).value('/x[2]', 'nvarchar(max)'), '') [new_2],
ISNULL(CAST(N'<x>' + REPLACE(Combined, '~', N'</x><x>') + N'</x>' AS XML).value('/x[3]', 'nvarchar(max)'), '') [new_3]
FROM CTE;