DAX - How to get RANKX independent of filters

[Anonymous]

// Note: [Master Product] is not the key level of Product but a higher level "Parent" product
// Countrows works perfectly and retains the same Total of Category value even when used within additional Product filters

measure 'Sales'[MyCount] = 
var tbProds = 
  calculatetable(filter(all('Product'[Master Product]),[Sales]>0), ALLEXCEPT('Product','Product'[Category]))
return 
  countrows(tbProds)
  
// Rankx only works as long as no additional filters are applied
// When additional Product filters are applied (e.g. Product[Vendor]) it ranks [Master Product] within that filter not within Total Category
// Why is one function affected by filters the other not?
// How do we achieve a ranking within Category that is independent of filters?

measure 'Sales'[MyRank] =
var tbProds = 
  calculatetable(filter(all('Product'[Master Product]),[Sales]>0), ALLEXCEPT('Product','Product'[Category]))
return 
  rankx(tbProds,[Sales],,,dense)

The above snippet shows 2 measures, MyCount works as expected, MyRank does not. Hopefully the comments tell most of the story. The goal is to get a rank over Product[Category] that is constant regardless of any other Product filters. The VARs are for readability only, we get the same results without using VARs. Any ideas how to achieve this? 

[v-chuncz-msft]

@Anonymous ,

You may try ALLSELECTED.

https://community.powerbi.com/t5/Desktop/RANK-MIN-OF-SUM-grouped-by-two-fields/m-p/581289#M275025

[Anonymous]

ALLSELECTED has the opposite effect to what I require. From https://docs.microsoft.com/en-us/dax/allselected-function-dax

"The ALLSELECTED function gets the context that represents all rows and columns in the query, while keeping explicit filters and contexts other than row and column filters."

I want to remove all filters except the current Product[Category] which should be what ALL() does.

I suspect this is something to do with RANKX being iterative but I dont understand it enough to explain why. I have even tried the countrows(filter(table,sales>currsales variable)) technique and that doesn't work either. Probably because of filter iterating too.

Thanks for replying though