Count dates and compare them at the same time SQL Server 2008 - sql-server-2008

Question

I have several datetime columns. I need to calculate in SQL Server 2008 for each timestamp how many datetime stamps in the same column are smaller than each of datetime stamps.
For example: for 2016-05-01 14:24:000.00 in column DateTime1 I need to calculate how many datetime values are smaller then it in DateTime1 column.
I also need to know how many datetimestamps are smaller than a datetime stamp for the same record (in the same row) in column DateTime2 and 3.
DateTime1 DateTime2 DateTime3
----------------------------------------------------------------------------
2016-05-01 13:24:000.00 2016-05-01 15:24:000.00 2016-05-01 16:20:000.00
2016-05-01 13:30:000.00 2016-05-01 14:21:000.00 2016-05-01 15:10:000.00
2016-05-01 14:24:000.00 2016-05-01 17:21:000.00 2016-05-01 18:10:000.00

Answer 1

If I understand correctly, you can use rank():
select t.*,
rank() over (order by datetime1) as dt1_rank,
rank() over (order by datetime2) as dt2_rank,
rank() over (order by datetime3) as dt3_rank
from t ;
Depending on how you want to treat tied values, you might actually want dense_rank(). Also, you might want to subtract 1 from the ranking value.

Answer 2

Assume I have a Table name [TestTB] has 3 columns DateTime1,DateTime2,DateTime3 .
I said CountSmallerDateTime1 as "how many datetime values are smaller then it in DateTime1 column"
I said CountSmallerDateTime2 as "how many datetimestamps are smaller than a datetime stamp for the same record (in the same row) in column DateTime2" , Similarity , CountSmallerDateTime3 for DateTime3 .
Then I have a query for your request :
SELECT [DateTime1]
,[DateTime2]
,[DateTime3]
,(SELECT COUNT(1)
FROM [TestTB] Sub
WHERE TB.[DateTime1] >Sub.[DateTime1]) AS CountSmallerDateTime1
,(
CASE WHEN TB.[DateTime2] > TB.[DateTime1] AND TB.[DateTime2] > TB.[DateTime3] THEN
2
WHEN ( (TB.[DateTime2] <= TB.[DateTime1] AND TB.[DateTime2] > TB.[DateTime3])
OR (TB.[DateTime2] > TB.[DateTime1] AND TB.[DateTime2] <= TB.[DateTime3])) THEN
1
ELSE
0
END
) AS CountSmallerDateTime2,
(
CASE WHEN TB.[DateTime3] > TB.[DateTime1] AND TB.[DateTime3] > TB.[DateTime2] THEN
2
WHEN ( (TB.[DateTime3] <= TB.[DateTime1] AND TB.[DateTime3] > TB.[DateTime2])
OR (TB.[DateTime3] > TB.[DateTime1] AND TB.[DateTime3] <= TB.[DateTime2])) THEN
1
ELSE
0
END
) AS CountSmallerDateTime3 FROM [TestTB] TB

Answer 3

;WITH CTE(DATE1, DATE2, DATE3,RN)
AS
(
SELECT CONVERT(DATETIME , '2016-05-01 13:24:000.00'), CONVERT(DATETIME,'2016-05-01 15:24:000.00'), CONVERT(DATETIME,'2016-05-01 16:20:000.00'),1
UNION ALL
SELECT CONVERT(DATETIME , '2016-05-01 13:30:000.00'), CONVERT(DATETIME,'2016-05-01 14:21:000.00'), CONVERT(DATETIME,'2016-05-01 15:10:000.00'),2
UNION ALL
SELECT CONVERT(DATETIME , '2016-05-01 14:24:000.00'), CONVERT(DATETIME,'2016-05-01 17:21:000.00'), CONVERT(DATETIME,'2016-05-01 18:10:000.00'),3
)
SELECT RANK() OVER (ORDER BY DATE1) -1 AS SAME_COLUMN_DATE1
, RANK() OVER (ORDER BY DATE2) -1 AS SAME_COLUMN_DATE2
, RANK() OVER (ORDER BY DATE3) -1 AS SAME_COLUMN_DATE3
, CASE WHEN RN=1 AND DATE1< DATE2 AND DATE1<DATE3 THEN 0
WHEN RN=1 AND DATE1< DATE2 AND DATE1>DATE3 THEN 1
WHEN RN=1 AND DATE1> DATE2 AND DATE1<DATE3 THEN 1
ELSE 2
SAME_ROW_1
, CASE WHEN RN=2 AND DATE2< DATE1 AND DATE2<DATE3 THEN 0
WHEN RN=2 AND DATE2< DATE1 AND DATE2>DATE3 THEN 1
WHEN RN=2 AND DATE2> DATE1 AND DATE2<DATE3 THEN 1
ELSE 2
END SAME_ROW_2
, CASE WHEN RN=3 AND DATE3< DATE1 AND DATE3<DATE2 THEN 0
WHEN RN=3 AND DATE3< DATE1 AND DATE3>DATE2 THEN 1
WHEN RN=3 AND DATE3> DATE1 AND DATE3<DATE2 THEN 1
ELSE 2
END SAME_ROW_3
FROM CTE ORDER BY RN