cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
Nick_M New Contributor
New Contributor

Re: Calculate Turn Around Time

Easiest thing to do is get a dedicated Date Table with a column that denotes workday or not.  Plus having a dedicated date table opens up all the time intelligence functions.

Panda2018 Member
Member

Re: Calculate Turn Around Time

Can you show me in an example hoew to?

Thanks, & how do I share an dataset?

 

@Nick_M Thanks!

Nick_M New Contributor
New Contributor

Re: Calculate Turn Around Time

Sure thing. I will use your first table as the starting point.  You can use excel to build a date table and then import, or use the CALENDAR function in DAX to create one.  I went with the DAX.

 

Calculations-->Modeling-->New Table.  Use the CALENDAR function.  Which needs a start and end date (which will cover in a sec).  Now, could use CALENDARAUTO which will find the start and end data automatically, but does so by searching the entire data model.  Would have worked fine in this small example, but in bigger data models that have date place holders as something like 1/1/9999, CALENDARAUTO will find that as the max.  So back to our CALENDAR function.  Need to find the Min and Max.  We wil do this by search each the Contact Date and Date writting columns and use the Min and Max of those two columns:

Date = 
CALENDAR(
	MIN (
		MIN (TurnAroundTIme[Contact Date]),
		MIN( TurnAroundTIme[Date Written])
	),
	MAX (
		MAX (TurnAroundTIme[Contact Date]),
		MAX( TurnAroundTIme[Date Written])
	)
)

Then add a calculated column for Day Name and then one to label using that day name to denote Weekday or Weekend:

Day = FORMAT('Date'[Date],"DDDD")

Day Type = 
SWITCH(
	'Date'[Day],
	"Saturday", "Weekend",
	"Sunday", "Weekend",
	"Weekday"
)

Here's the end result:

Date Table.png

 

Then need to mark as  Date Table so DAX knows to use that for the built-in time intelligence functions:

Mark as date table.png

Now that we have a calendar that we can use, setting up the DAX formulas should be much easier.   There's actually two ways we can do this, one using FILTER and one using DATESBETWEEN.  

TaT using Filter = 
CALCULATE(
    COUNTROWS( 'Date' ),
        FILTER( ALL('Date'),
		    MAX( TurnAroundTime[Contact Date]) <= 'Date'[Date]
		    && MAX( TurnAroundTime[Date Written]) >= 'Date'[Date]
            && 'Date'[Day Type] ="Weekday"
	    )
)

TaT using DatesBetween =
 CALCULATE(
	COUNTROWS( 'Date' ),
       DATESBETWEEN('Date'[Date],
		MAX( TurnAroundTime[Contact Date]) ,
		MAX( TurnAroundTime[Date Written])
    ),
    'Date'[Day Type] ="Weekday"
)

DATESBETWEEN leverages the use of the time-intelligence, but requires some additional logic to not give a figure when there is no Contact Date or Date Written field:

TaT using DatesBetween = 
IF (
    NOT
        OR(
         ISBLANK(MAX( TurnAroundTime[Contact Date])), 
         ISBLANK(MAX( TurnAroundTime[Date Written])
         )
    ),
    CALCULATE(
	    COUNTROWS( 'Date' ),
          DATESBETWEEN('Date'[Date],
	    	MAX( TurnAroundTime[Contact Date]) ,
	    	MAX( TurnAroundTime[Date Written])
      	),
    'Date'[Day Type] ="Weekday"
    )
 )

Then the final output:

Final Table.png

sailkitty Regular Visitor
Regular Visitor

Re: Calculate Turn Around Time

For the least amount of typing do this in the Power Query Editor. 

Select the two Columns- 

Go tot he Add Column Ribon at the top and slecet the Date drop down in "From Date & Time"  and click "Subtract Days" 

 

 

Hope that helps 

Panda2018 Member
Member

Re: Calculate Turn Around Time

Thank you @sailkitty

It works...But for same date it should say 1 day not 0

Also it shows numbers in negative like difference between 11/1/18 & 11/2/18 is showing -2.

 

Thanks,

panda2018

sailkitty Regular Visitor
Regular Visitor

Re: Calculate Turn Around Time

Part of the benefit of doing it that way is it shows you negative digits where as the Measures will result in an error. This allows you to filter them out later (in  measures ) or correct your underlying data that has errors (for example when items are "delivered" before the order is place. This is a data error) 

 

Esp where a blank might default to 01/01/1900 

Panda2018 Member
Member

Re: Calculate Turn Around Time

Thank you so very much!!! @Nick_M

 

-panda2018