SQL Server : DATEDIFF exclude certain time within the two dates - sql

Question

I'm trying to get the difference between two dates in minutes, is it possible for me to exclude lunch times from this calculation aka 13:30-14:00 purely within my SQL statement. What I'm currently doing is below:
Example row
SetupStart StartTime SetupTime
----------------------------------------------------------------
2017-01-23 12:56:42.000 2017-01-23 14:41:06.000 105
My current statement:
DATEDIFF(MINUTE, UserData.Orders.SetupStart, UserData.Orders.StartTime) AS[SetupTime]
EDIT:--------------
Here is my current Select statement
SELECT DISTINCT
TOP (100) PERCENT UserData.Resources.Name AS Resource, UserData.Orders.OrderNo, UserData.Orders.StringAttribute4 AS Customer, UserData.Orders.Product,
CEILING(UserData.Orders.OpNo10Quantity) AS OpNo10Quantity, UserData.Orders.NumericalAttribute12 AS Speed, UserData.Orders.SetupStart, UserData.Orders.StartTime,
UserData.Orders.EndTime, CASE ToggleAttribute1 WHEN 1 THEN 'Yes' ELSE 'No' END AS Packed, UserData.Orders.StringAttribute5 AS OrderInstructions, UserData.Orders.NextResource,
UserData.Orders.DatasetId, UserData.Orders_Dataset.name AS ScheduleName, UserData.Orders.ShowOnReport, dbo.tblPreactorExportFullv10.Length AS L,
dbo.tblPreactorExportFullv10.Thickness AS T, dbo.tblPreactorExportFullv10.fWidth AS W,
DATEDIFF(MINUTE, UserData.Orders.SetupStart, UserData.Orders.StartTime) AS [SAM SetupTime]
FROM UserData.Orders INNER JOIN
UserData.Resources ON UserData.Orders.Resource = UserData.Resources.ResourcesId AND UserData.Orders.Resource = UserData.Resources.ResourcesId INNER JOIN
UserData.Orders_Dataset ON UserData.Orders.DatasetId = UserData.Orders_Dataset.DatasetId AND UserData.Orders.DatasetId = UserData.Orders_Dataset.DatasetId INNER JOIN
dbo.tblPreactorExportFullv10 ON UserData.Orders.PartNo = dbo.tblPreactorExportFullv10.ProductCode
WHERE (UserData.Orders.DatasetId = 15) AND (UserData.Resources.Name = 'Moulder 6') AND (UserData.Orders.ShowOnReport = 1)
AND (UserData.Orders.OperationProgress <> 5)
ORDER BY UserData.Orders.SetupStart

Answer 1

I'm sure Gordon's answer will be faster, but another option is as follows:
(Edit: Just for fun, ran this on 20,000 records and returned in 238 milliseconds)
Declare #YourTable table (SetupStart datetime, StartTime datetime)
Insert Into #YourTable values
('2017-01-23 12:56:42.000','2017-01-23 14:41:06.000'),
('2017-01-23 15:00:00.000','2017-01-23 18:30:00.000'), -- No Lunch
('2017-01-23 23:51:00.000','2017-01-23 23:53:46.000'), -- Anomoly mentioned
('2017-01-23 13:15:00.000','2017-01-23 13:45:00.000') -- Started After Lunch
Select A.*
,B.*
From ( ... your complex query here ... ) A
Cross Apply (
Select SetupTime = count(*)
From (Select Top (DateDiff(MINUTE,A.SetupStart,A.StartTime)) T=cast(DateAdd(MINUTE,Row_Number() Over (Order By Number)-1,A.SetupStart) as time)
From master..spt_values ) S
Where (cast(A.SetupStart as time)<'13:30' and cast(A.StartTime as time)>'14:00' and T not between '13:30' and '14:00')
or (cast(A.SetupStart as time) > '13:30' )
or (cast(A.StartTime as time) < '14:00' )
) B
Returns
SetupStart StartTime SetupTime
2017-01-23 12:56:42.000 2017-01-23 14:41:06.000 75
2017-01-23 15:00:00.000 2017-01-23 18:30:00.000 210
2017-01-23 23:51:00.000 2017-01-23 23:53:46.000 2
2017-01-23 13:15:00.000 2017-01-23 13:45:00.000 30

Answer 2

declare
#ldt_from time = '13:00'
,#ldt_to time = '14:00';
-- dummy data prepare
;with [my_setups] as
(
select [SetupStart] = cast('2017-01-23 13:56:42.000' as datetime), [StartTime] = cast('2017-01-23 14:41:06.000' as datetime)
union all
select [SetupStart] = cast('2017-01-23 12:45:00.000' as datetime), [StartTime] = cast('2017-01-23 12:46:00.000' as datetime)
)
-- end dummy data prepare
select [minutes] =
datediff(mi, [SetupStart], [StartTime])
- datediff(day, [SetupStart], [StartTime]) * (datediff(mi, #ldt_from, #ldt_to))
+ case when cast([SetupStart] as time) between #ldt_from and #ldt_to or cast([StartTime] as time) between #ldt_from and #ldt_to then
datediff
(
mi
,case when cast([SetupStart] as time) > #ldt_from then cast([SetupStart] as time) else #ldt_from end
,case when cast([StartTime] as time) < #ldt_to then cast([StartTime] as time) else #ldt_to end
)
else 0 end
from
[my_setups]

Answer 3

This is a pain in SQL Server, but you can do it:
(CASE WHEN CAST(o.SetupStart as time) > '14:00:00' OR
CAST(o.StartTime as time) < '13:30:00'
THEN DATEDIFF(MINUTE, o.SetupStart, o.StartTime)
WHEN CAST(o.SetupStart as time) >= '13:30:00' AND
CAST(o.StartTime as time) <= '14:00:00'
THEN 0
WHEN CAST(o.SetupStart as time) >= '13:30:00'
THEN DATEDIFF(MINUTE,
CAST(CAST(o.SetupStart as DATE) as DATETIME) + CAST('14:00:00' as TIME),
o.StartTime
)
WHEN CAST(o.StartTime as Time) <= '14:00:00'
THEN DATEDIFF(MINUTE,
o.SetStart,
CAST(CAST(o.StartTime as DATE) as DATETIME) + CAST('13:30:00' as TIME)
)
ELSE DATEDIFF(MINUTE, o.SetupStart, o.StartTime) - 30
END) AS [SetupTime]
This handles the following cases:
Times are both before or after lunch.
Times are both during lunch.
SetupStart is during lunch.
StartTime is during lunch.
SetupStart is before lunch and StartTime is after lunch

Answer 4

The easiest way is what #Shakeer Hussain recommended
DATEDIFF(MINUTE, UserData.Orders.SetupStart, UserData.Orders.StartTime) AS[SetupTime]
BECOMES
DATEDIFF(MINUTE, UserData.Orders.SetupStart, DATEADD(MINUTE,-30,(UserData.Orders.StartTime))) AS[SetupTime]
Additionally, you could do two DATEDIFF functions and add them together:
DATEDIFF(MINUTE, UserData.Orders.SetupStart, UserData.Orders.LunchStartTime)
+ DATEDIFF(MINUTE, UserData.Orders.LunchEndTime, UserData.Orders.StartTime) AS[SetupTime]
If these data points don't exist, you can create a temp table.