how to separated One column to two with cases - sql

Question

I have a table and I would want to separate the data to multiple columns, how i can do it ?
I tried this:
select a.[batch],a.[Doc_Type],
Soaking Out =
CASE a.[Doc_Type]
WHEN 'BB' THEN 'Soaking Out'
END,
Soaking In =
CASE a.[Doc_Type]
WHEN 'AA' THEN 'Soaking In'
END,
FROM Transaction_Hdr a JOIN Transaction_dtl b
on a.Doc_Number=b.Doc_Number

Answer 1

Your original query would output the strings 'soaking in' or 'soaking out', but what is needed in those case expressions (after then) is the column [Qty] and it is that value which will be returned from the case expression.
What I don't know is which table [Qty] comes from but I assume it is the detail table (b) otherwise there isn't much point in joining that detail table.
SELECT
a.[Doc_Type]
, a.[batch]
, CASE a.[Doc_Type] WHEN 'BB' THEN b.Qty END [soaking out]
, CASE a.[Doc_Type] WHEN 'AA' THEN b.Qty END [soaking in]
FROM Transaction_Hdr a
JOIN Transaction_dtl b ON a.Doc_Number = b.Doc_Number
ORDER BY
a.[Doc_Type]
, a.[batch]
But: a "detail" table and a "header" table usually indicates many rows of detail for a single header. So you might need a SUM() and GROUP BY
SELECT
h.[Doc_Type]
, h.[batch]
, SUM(CASE h.[Doc_Type] WHEN 'BB' THEN d.Qty END) [soaking out]
, SUM(CASE h.[Doc_Type] WHEN 'AA' THEN d.Qty END) [soaking in]
FROM Transaction_Hdr h
JOIN Transaction_dtl d ON h.Doc_Number = d.Doc_Number
GROUP BY
h.[Doc_Type]
, h.[batch]
ORDER BY
h.[Doc_Type]
, h.[batch]
Note I have now used aliases "h" = "header" and "d" = "detail" as I am really not keen of aliases that rely on a sequence within the query (as that sequence can get messed with very easily). I find it way easier for an alias to easily identify its associated table by "first letter of each word in a table's name" or similar.

Answer 2

select a.[batch],a.[Doc_Type],
isnull(CASE WHEN a.[Doc_Type]='AA' THEN convert(real,a.Qty) END,0) as [Soaking In] ,
isnull(CASE WHEN a.[Doc_Type]='BB' THEN convert(real,a.Qty) END ,0)as [Soaking Out]
FROM Transaction_Hdr a

Answer 3

I think you are looking for quantity in result table, So you should use that instead of string 'Soaking In' and 'Soaking Out' as follows
select a.[batch],a.[Doc_Type],
SoakingOut =
CASE a.[Doc_Type]
WHEN 'BB' THEN Qty
END ,
SoakingIn =
CASE a.[Doc_Type]
WHEN 'AA' THEN Qty
END
FROM #temp a

Answer 4

BEGIN TRAN
CREATE TABLE #Data (
Doc_Type VARCHAR(10),
Batch INT,
Qty DECIMAL(4,2)
);
INSERT INTO #Data VALUES
('AA', 1, 20.5),
('BB', 2, 10 ),
('AA', 3, 6 ),
('BB', 4, 7 ),
('AA', 5, 8 );
SELECT ISNULL(CASE WHEN Doc_Type='AA'THEN CONVERT(NVARCHAR(10),QTY) END,'') Soaking_In ,
ISNULL(CASE WHEN Doc_Type='BB'THEN CONVERT(NVARCHAR(10),QTY) END,'') Soaking_Out
FROM #Data
ROLLBACK TRAN

Answer 5

Use CASE() and Modulus as below, assuming that Batch is always inceremnted by 1 and Doc_Type has always those two values AA and BB in the same order:
CREATE TABLE Data (
Doc_Type VARCHAR(10),
Batch INT,
Qty DECIMAL(4,2)
);
INSERT INTO Data VALUES
('AA', 1, 20.5),
('BB', 2, 10 ),
('AA', 3, 6 ),
('BB', 4, 7 ),
('AA', 5, 8 );
SELECT D.Doc_Type, D.Batch,
CASE WHEN D.Batch % 2 = 0 Then 0 ELSE D.Qty END AS Soaking_In,
CASE WHEN D.Batch % 2 = 1 Then 0 ELSE D.Qty END AS Soaking_Out
FROM Data D;
Results:
+----------+-------+------------+-------------+
| Doc_Type | Batch | Soaking_In | Soaking_Out |
+----------+-------+------------+-------------+
| AA | 1 | 20,50 | 0,00 |
| BB | 2 | 0,00 | 10,00 |
| AA | 3 | 6,00 | 0,00 |
| BB | 4 | 0,00 | 7,00 |
| AA | 5 | 8,00 | 0,00 |
+----------+-------+------------+-------------+
Demo