Transform view of table using sql - sql

Question

I have the following table
Date Input Output Rate
10/10/2001 10 15 0.5
10/11/2002 13 14 0.3
1. Date Input Output Rate
2. 10/10/2001 10 15 0.5
3. 10/11/2002 13 14 0.3
I need to transform this table in the following table:
1. Date Category Amount
2. 10/10/2001 Input 10
3. 10/10/2001 Output 15
4. 10/10/2001 Rate 0.5
5. 10/11/2002 Input 13
6. 10/11/2002 Output 14
7. 10/11/2002 Rate 0.3
Which query should I use?
Thank you in advance.

Answer 1

You can use UNPIVOT:
DECLARE #T TABLE (Date DATE, Input DECIMAL(5, 1), [Output] DECIMAL(5, 1), Rate DECIMAL(5, 1));
INSERT #T VALUES ('20011010', 10, 15, 0.5), ('20021110', 13, 14, 0.3);
SELECT upvt.Date, upvt.Category, upvt.Amount
FROM #T
UNPIVOT
( Amount
FOR Category IN ([input], [Output], [Rate])
) upvt;
However, I prefer the slightly more flexible CROSS APPLY with a table valued constructor:
DECLARE #T TABLE (Date DATE, Input DECIMAL(5, 1), [Output] DECIMAL(5, 1), Rate DECIMAL(5, 1));
INSERT #T VALUES ('20011010', 10, 15, 0.5), ('20021110', 13, 14, 0.3);
SELECT t.Date, upvt.Category, upvt.Amount
FROM #T AS t
CROSS APPLY
(VALUES
('Input', t.Input),
('Output', t.[Output]),
('Rate', t.Rate)
) AS upvt (Category, Amount);
As it affords you more flexibility with your datatype conversions, and with renaming your columns slightly differently, e.g. it is easier to make a column like CallStartTime something like Call Start, so it is more report friendly.
e.g if your input and output columns are actually integers, you will get an error message when using UNPIVOT:
The type of column "Rate" conflicts with the type of other columns specified in the UNPIVOT list.
So you would need to use a subquery, and convert your columns there:
DECLARE #T TABLE (Date DATE, Input INT, [Output] INT, Rate DECIMAL(5, 1));
INSERT #T VALUES ('20011010', 10, 15, 0.5), ('20021110', 13, 14, 0.3);
SELECT upvt.Date, upvt.Category, upvt.Amount
FROM ( SELECT Date,
Input = CONVERT(DECIMAL(5, 1), Input),
[Output] = CONVERT(DECIMAL(5, 1), [Output]),
Rate
FROM #T
) AS t
UNPIVOT
( Amount
FOR Category IN ([input], [Output], [Rate])
) upvt;
And while you have to do the same conversion with CROSS APPLY, it seems like less clutter to me:
DECLARE #T TABLE (Date DATE, Input INT, [Output] INT, Rate DECIMAL(5, 1));
INSERT #T VALUES ('20011010', 10, 15, 0.5), ('20021110', 13, 14, 0.3);
SELECT t.Date, upvt.Category, upvt.Amount
FROM #T AS t
CROSS APPLY
(VALUES
('Input', CONVERT(DECIMAL(5, 1), Input)),
('Output', CONVERT(DECIMAL(5, 1), [Output])),
('Rate', t.Rate)
) AS upvt (Category, Amount);
It is of course, personal preference.

Answer 2

You might use a simple Standard SQL UNION ALL which is supported by every DBMS -> portable :-)
SELECT Date, 'Input' AS Category, Cast(Input as DEC(5,1)) AS Amount
FROM Tab
UNION ALL
SELECT Date, 'Output' AS Category, Output
FROM Tab
UNION ALL
SELECT Date, 'Rate' AS Category, Rate
FROM Tab