MAX Date of multiple columns? - sql

Question

How do you return 1 value per row of the max of several columns:
TableName [RefNumber, FirstVisitedDate, SecondVisitedDate, RecoveryDate, ActionDate]
I want MaxDate of (FirstVisitedDate, SecondVisitedDate, RecoveryDate, ActionDate) these dates for all rows in single column and I want another new column(Acion) depends on Max date column for ex: if Max date is from FirstVisitedDate then it will be 'FirstVisited' or if Max date is from SecondVisitedDate then it will be 'SecondVisited'...
The Total result Like:
Select RefNumber, Maxdate, Action From Table
group by RefNumber

Answer 1

The brute force approach isn't so bad with four columns:
select (case when FirstVisitedDate >= SecondVisitedDate and
FirstVisitedDate >= RecoveryDate and
FirstVisitedDate >= ActionDate
then FirstVisitedDate
when SecondVisitedDate >= RecoveryDate and
SecondVisitedDate >= ActionDate
then SecondVisitedDate
when RecoveryDate >= ActionDate
then RecoveryDate
else ActionDate
end),
(case when FirstVisitedDate >= SecondVisitedDate and
FirstVisitedDate >= RecoveryDate and
FirstVisitedDate >= ActionDate
then 'FirstVisitedDate'
when SecondVisitedDate >= RecoveryDate and
SecondVisitedDate >= ActionDate
then 'SecondVisitedDate'
when RecoveryDate >= ActionDate
then 'RecoveryDate'
else 'ActionDate'
end)
from table t;
EDIT:
Doing this in a group by is just a question of adding aggregation functions:
select RefNumber,
(case when max(FirstVisitedDate) >= max(SecondVisitedDate) and
max(FirstVisitedDate) >= max(RecoveryDate) and
max(FirstVisitedDate) >= max(ActionDate)
then max(FirstVisitedDate
when max(SecondVisitedDate) >= max(RecoveryDate) and
max(SecondVisitedDate) >= max(ActionDate)
then max(SecondVisitedDate)
when max(RecoveryDate) >= max(ActionDate)
then max(RecoveryDate)
else max(ActionDate)
end),
(case when max(FirstVisitedDate) >= max(SecondVisitedDate) and
max(FirstVisitedDate) >= max(RecoveryDate) and
max(FirstVisitedDate) >= max(ActionDate)
then 'FirstVisitedDate'
when max(SecondVisitedDate) >= max(RecoveryDate) and
max(SecondVisitedDate) >= max(ActionDate)
then 'SecondVisitedDate'
when max(RecoveryDate) >= max(ActionDate)
then 'RecoveryDate'
else 'ActionDate'
end)
from table t
group by RefNumber;

Answer 2

;WITH cte AS (
-- Build table of date
SELECT [RefNumber],
[ActionDate]= [FirstVisitedDate],
[Action] = 'First Visited'
FROM [Table1]
UNION ALL
SELECT [RefNumber],
[SecondVisitedDate],
'Second Visited'
FROM [Table1]
UNION ALL
SELECT [RefNumber],
[RecoveryDate],
'Recover'
FROM [Table1]
), cte2 AS (
-- Add row_number to pull most recent to top
SELECT [RefNumber],
[Action],
[ActionDate],
[DateRank] =
ROW_NUMBER() OVER (PARTITION BY RefNumber
ORDER BY [ActionDate] DESC)
FROM [cte]
)
-- select only the most recent
SELECT *
FROM cte2
WHERE [DateRank] = 1

Answer 3

SELECT RecordID, MaxDate
FROM SourceTable
CROSS APPLY (SELECT MAX(d) MaxDate
FROM (VALUES (date1), (date2), (date3),
(date4), (date5), (date6),
(date7)) AS dates(d)) md

Answer 4

I wrote a custom function to do this:
CREATE FUNCTION [dbo].[MaxOf5]
(
#D1 DateTime,
#D2 DateTime,
#D3 DateTime,
#D4 DateTime,
#D5 DateTime
)
RETURNS DateTime
AS
BEGIN
DECLARE #Result DateTime
SET #Result = COALESCE(#D1, #D2, #D3, #D4, #D5)
IF #D2 IS NOT NULL AND #D2 > #Result SET #Result = #D2
IF #D3 IS NOT NULL AND #D3 > #Result SET #Result = #D3
IF #D4 IS NOT NULL AND #D4 > #Result SET #Result = #D4
IF #D5 IS NOT NULL AND #D5 > #Result SET #Result = #D5
RETURN #Result
END
To call this, and calculated your Action column, this should work:
SELECT
MaxDate,
CASE WHEN MaxDate = FirstVisitedDate THEN 'FirstVisited'
WHEN MaxDate = SecondVisitedDate THEN 'SecondVisited'
WHEN MaxDate = RecoveryDate THEN 'Recovery'
WHEN MaxDate = ActionDate THEN 'Action'
END AS [Action]
FROM (
SELECT
RefNumber,
dbo.MaxOf5(FirstVisitedDate, SecondVisitedDate, RecoveryDate, ActionDate) AS MaxDate,
FirstVisitedDate, SecondVisitedDate, RecoveryDate, ActionDate
FROM table
) AS data
Note that is is possible for more than one of your dates to be tied for the max date. In this case, the order of your WHEN clauses determines which one wins.