Ranking by combination of two columns

[RobGer]

Dear PowerBI Community,

I have the following question, hopefully you can help me with this.

I have a table for each product sold. It looks like this:

Product	Region	Main Seller	Support Seller
1	West	Alex	Pedro
2	East	Stephanie	Alex
3	North	Pedro
4	North	Stephanie	Pedro
5	South	Pedro	Alex
6	South	Pedro	Alex
7	South	Alex	Stephanie

I would like to calculate an rank the top seller combinations, independent of if they supported during the sale, or if they were the main seller. 

The desired result would be displayed on a table with a ranking. 

In this example, the combination "Alex and Pedro" would appear 3 times, and ranked as first. "Stephanie and Alex" would be ranked second, and the rest would follow with only one count ("Stephanie and Alex").

I don´t wish to display the sales, where there is no support seller, but I believe I can filter that out later through the filter pane and setting "Support Seller" to not allowing blanks.

Thank you for your help in advance!

Regards,

RobGer

[ThxAlot]

Interesting, this seemly easy question turns out to be brain teaser.

Rnk = 
RANKX(
    ALLSELECTED( PAIR[Product] ),
    VAR __prod = CALCULATETABLE( VALUES( PAIR[Product] ), ALLEXCEPT( PAIR, PAIR[Product] ) )
    VAR __main = CALCULATE( MAX( PAIR[Main Seller] ), ALL(), __prod )
    VAR __sup = CALCULATE( MAX( PAIR[Support Seller] ), ALL(), __prod )
    RETURN
        CALCULATE(
            COUNTROWS( PAIR ),
            TREATAS( { ( __main, __sup ) }, PAIR[Main Seller], PAIR[Support Seller] )
        )
            + CALCULATE(
                COUNTROWS( PAIR ),
                ALL( PAIR[Product] ),
                TREATAS( { ( __sup, __main ) }, PAIR[Main Seller], PAIR[Support Seller] )
            ),
    COUNTROWS( PAIR )
        + CALCULATE(
            COUNTROWS( PAIR ),
            TREATAS(
                { ( MAX( PAIR[Support Seller] ), MAX( PAIR[Main Seller] ) ) },
                PAIR[Main Seller],
                PAIR[Support Seller]
            )
        ),
    ,
    DENSE
)

[Ahmedx]

to get a suitable model you need to create tables in Power Query.
please look at the file and study it

[olgad]

Hi, 

CombinedName =

IF(

    'Sales'[Main] <= 'Sales'[Supporter],

    'Sales'[Main] & " " & 'Sales'[Supporter],

    'Sales'[Supporter] & " " & 'Sales'[Main]

)

[amustafa · Accepted solution]

Stange request, anyhow...change the Seller Combination as following.

Seller Combination =

IF(

    SalesTable[Main Seller] < SalesTable[Support Seller],

    SalesTable[Main Seller] & ", " & SalesTable[Support Seller],

    SalesTable[Support Seller] & ", " & SalesTable[Main Seller]

)

[amustafa]

Hi @RobGer 

Based on your sample dataset, I would create 1 calculated column and 2 measures.

Seller Combination = SalesTable[Main Seller] & " and " & SalesTable[Support Seller]

Seller Combination Count =

CALCULATE(

    COUNTROWS(SalesTable),

    ALLEXCEPT(SalesTable, SalesTable[Seller Combination])

)

Seller Combination Rank =

RANKX(

    ALL(SalesTable[Seller Combination]),

    [Seller Combination Count],

    ,DESC,

    Dense

)

Results:

If I answered your question, please mark this thread as accepted.
Follow me on LinkedIn:
https://www.linkedin.com/in/mustafa-ali-70133451/

[RobGer]

Hi @amustafa ,

this is looking good, I would need the combinations to be added together, in this case, "Pedro and Alex" would count the same as "Alex and Pedro". I don´t know if there is a feasible solution that would allow it.

[olgad]

Hi, 

Rank =

RANKX(

    CROSSJOIN(

        ALL(Sales[Main]),

        ALL(Sales[Supporter])),

        Sales[SalesAmount]

)

[RobGer]

Hi @olgad ,

This is looking in the right direction, would it be possible to fuse together the repeated seller combinations? In your example, "Alex and Dana" and "Dana and Alex" would be displayed together (as either "Alex and Dana" or "Dana and Alex") and the "Sales Amount" would be added together, too.