Cumulative % using Index not dates in DAX without measures

[Christian_DC]

Hello Everybody,

Thanks for taking the time to read my question.

I am trying to assign a cumulative % to each row of customer data based on descending values, so that I can then add another column to label each customer 'high value' (top 80% by value), 'medium value' (next 15%) or 'low value' (final 5%).  I have tried to create this using DAX as follows:

(In case the above picture hasn't come through properly, the column formula looks like this:)

Running % =
CALCULATE (
SUM ( 'Customer level'[Value %] ),
FILTER (
ALL ( 'Customer level'[Index.1] ,
'Customer level'[Index.1] <= MAX ( 'Customer level'[Index.1] )
)
)
)

I have added an Index column to give a rank but the above formula does not work.

Please help!

🙂

[TomMartens]

Hey,

I think if you rewrite your calculated column in this way, it should work

Running % =
var currentIndex =  'Customer level'[Index.1]
return
CALCULATE (
SUM ( 'Customer level'[Value %] ),
FILTER (
  ALLEXCEPT('Customer level', 'Customer    level'[nameofyourcustomercolumn]) ,
'Customer level'[Index.1] <= currentIndex 
)
)

Hope this helps, if not please provide some sample data.

Regards

[Christian_DC · Accepted solution]

Thanks Tom.

I actually just got it to work by modifying my original column formula to, i.e. using 'EARLIER' instead of 'MAX' :

Running % =
CALCULATE (
    SUM ( 'Customer level'[Value %] ),
    ALL ( 'Customer level' ),
    'Customer level'[Index.1] <= EARLIER ( 'Customer level'[Index.1] )
)

Incidentally, I have subsequently tried your method but haven't quite got it to work, but I won't trouble you further.

Many thanks for your quick reply!

Christian

[TomMartens]

Hey,

great you figured it out, but I would be a little hesitant to use

FILTER(ALL('table'), ...)

in my little sample data this leads to this result

Edited just realized that the column "Using Earlier" should be named "Using ALL(table)"

But maybe I was wrong in understanding your question not to create a cumulative sum for each customer separately.

Regards