Multiple date filters in calculatetable

[Anonymous]

Hi,
I saw this code snippet in a wokrshop held by Jeffrey Wang. The snippet defines a measure to calculate the number of returning customers  by date, in a very standard data model with sales, products and customers.

-----------------------

# NEW customers =

VAR vCustomersThisPeriod= VALUES(Appended_actuals[CustomerID])

var vCurrentdate = MIN('Date'[Date])

var VCustomersPreviousPeriod = CALCULATETABLE(VALUES(Sales[CustomerID]),
FILTER(ALL('Date'[Date]),'Date'[Date]<vCurrentdate),ALL('Date'[Date]))

Return COUNTROWS(EXCEPT(vCustomersThisPeriod,VCustomersPreviousPeriod))

-----------------------

My question is , why is there a need for the second ALL('Date'[Date]) in VCustomersPreviousPeriod?
Doesn't the first filter already accomplish the task?
Thanks a lot for the answers.

[Anonymous · Accepted solution]

Good spot! This ALL(...) is not necessary. It serves no purpose, even though it's not harmful, either. The FILTER overwrites any filter present on the column anyway, so removing any filters via ALL is redundant.

[Anonymous]

Hey daxer, thanks for the explanation!

[Anonymous]

Hi @amitchandak , thanks a lot for taking the time to answer, but I am not sure it responds to the question i had. In any case, it is much appreciated.

[amitchandak]

@Anonymous , I usually follow a measure approch

Current = sum(Table[Sales])

past =

Var _min = minx(allselected('Date'), Date[Date])

return

calculate(sum(Table[Sales]), filter(all('Date'), Date[Date] <_min)

new = countx(values(customer[Customer]), if(not(isblank([Current]))  && isblank([Past]) , [Customer], bank()))