Calculated column - creating partition based on multiple values

[tomislav_mi]

Hello all,

I am struggling with how to create one complex calculated column. Desired calculated column is in orange Active Channel Customer.

The idea is to use the column Active Channel Account to create a new column Active Channel Customer
- If the account is active in a specific month, then the Active Channel Account will be a channel for customer level - column Active Channel Customer
- If there are 2 types of channels then the value in the calculated column is going to be "both"

Report Date	Account	Customer	Active	Active Channel Account	Active Channel Customer
6/30/2022	a1	A	FALSE		Reseller
6/30/2022	a2	A	TRUE	Reseller	Reseller
6/30/2022	a3	A	FALSE		Reseller
6/30/2022	a4	A	TRUE	Reseller	Reseller
7/31/2022	a1	A	FALSE		Direct
7/31/2022	a2	A	FALSE		Direct
7/31/2022	a3	A	TRUE	Direct	Direct
7/31/2022	a4	A	FALSE		Direct
6/30/2022	b1	b	TRUE	Direct	Both
6/30/2022	b2	b	TRUE	Reseller	Both

Any help or advice is welcome.

Thanks!

[amitchandak · Accepted solution]

@tomislav_mi , Try a new column like

new column =
var _1 = countx(filter(Table, [Customer] = earlier([Customer]) && [Report Date] = earlier([Report Date]) && [Active Channel] = "Reseller"), [Customer])
var _2 = countx(filter(Table, [Customer] = earlier([Customer]) && [Report Date] = earlier([Report Date]) && [Active Channel] = "Direct"), [Customer])
return
Switch(true(),
not(isblank(_1)) && not(isblank(_2)) , "Both",
not(isblank(_1)),"Reseller",
"Direct"
)

[tomislav_mi]

@amitchandak 

Exactly what I was struggling to get! Great approach.
Thank you!