SQL: Showing column name of calculated result - sql

Question

So I am calculating the max of a series of column and I was wondering what the best way to show the name of the result column.
Example Table:
hour1 hour2 hour3
16 10 5
My query looks like this:
(SELECT Max(v)
FROM (VALUES (hour1) , (hour2) , (hour3))
AS VALUE (v)) AS PEAK_VALUE
Note this is in another query.
Desired output:
PEAK_VALUE PEAK_HOUR
16 hour1
I would also like to do further calculations on the PEAK_VALUE column as well. For example dividing it by 2 for this output:
PEAK_VALUE HALF_VALUE PEAK_HOUR
16 8 hour1

Answer 1

You almost got it though you have some issue with your query syntax. You need to add the column name to the un-pivot. After that user row_number() to find the max value
SELECT PEAK_VALUE = v,
HALF_VALUE = v / 2,
PEAK_HOUR = c
FROM
(
SELECT *, rn = row_number() over (order by v desc)
FROM example
CROSS APPLY
(
VALUES ('hour1', hour1) , ('hour2', hour2) , ('hour3', hour3)
) AS VALUE (c, v)
) v
WHERE rn = 1
dbfiddle

Answer 2

Just another option sans row_number()
Example
Select top 1
Peak_Value = Hour
,Half_Value = Hour/2
,Peak_Hour = Item
From YourTable A
Cross Apply (values ('hour1',hour1)
,('hour2',hour2)
,('hour3',hour3)
) B (Item,Hour)
Order by Hour Desc
Results
Peak_Value Half_Value Peak_Hour
16 8 hour1

Answer 3

Another approach using UNPIVOT.
DECLARE #table table(hour1 int, hour2 int, hour3 int)
insert into #table
values(16,10,5)
SELECT top 1 max(val) as peak_value, max(val) /2 as Half_Value, [hour]
FROM #table
unpivot (val for [hour] in ([hour1],[hour2],[hour3])) as t
group by [hour]
order by max(val) desc
peak_value
Half_Value
hour
16
8
hour1

Answer 4

You can shove a whole unpivoting query inside a CROSS APPLY, and then do further calculations:
SELECT t.*, v.*
FROM yourTable t
CROSS APPLY (
SELECT TOP (1)
v.PEAK_VALUE,
HALF_VALUE = v.PEAK_VALUE / 2,
v.PEAK_HOUR,
FROM (VALUES
('hour1', hour1),
('hour2', hour2),
('hour3', hour3)
) AS v(PEAK_HOUR, PEAK_VALUE)
ORDER BY PEAK_VALUE DESC
) AS v
This is slightly different from the other answers, in that it will calculate the subquery per row