RankX to show category and not ranking

[Tevon713]

Hi.

I can't seem to find an answer, simple question how would I show ranking of category over select period.

Right now, for the excerise I have it as follow:

Ranking Fruits = RANKX(ALL(Master[FruitName]),[Sum Value],,DESC,Dense)

Rank	FruitName	Sum Value
1	Apple	500
2	Pear	439
3	Banana	350
4	Oranges	299
5	Kiwi	100

I want to be able to show the ranking in matrix with ranking in 1st column, the time period as column and for each time period it to display the category and not the ranking. Something like the below. Thanks.

Rank	Jan-21	Feb-21	Mar-21
1	Apple	Banana	Pear
2	Pear	Apple	Oranges
3	Banana	Pear	Kiwi
4	Oranges	Oranges	Apple
5	Kiwi	Kiwi	Banana

[Ashish_Mathur]

Hi,

I am not sure of how much i can help, but i would like to try.  Share the link from where i can download your PBI file.

[Greg_Deckler · Accepted solution]

@Tevon713 OK, I did this with a disconnected Rankings table (just 1 - 5 in a single column called Rank) and then a "FruitsBasket" table. The measure is:

Fruit Rank = 
    VAR __Rank = MAX('Rankings'[Rank])
    VAR __Table = 
        SUMMARIZE(
            'FruitsBasket',
            [FruitName],
            "__Value",SUM([Value])
        )
    VAR __RankTable = 
        ADDCOLUMNS(
            __Table,
            "__Rank",RANKX(__Table,[__Value],,DESC)
        )
RETURN
    MAXX(FILTER(__RankTable,[__Rank] = __Rank),[FruitName])

This was my FruitsBasket table:

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WciwoyElV0lHySszTNTIEMkwNDJRidRASbqlJEAkTNAnfxCKIhBFUIiA1sQjZJBNUcbhBxqjicHNgNjsl5gEhsknG6DJwszD0wE0zhMr4FyXmpacWIxtnhCEFNw9TCm4gzDvemeWZyKYZoorDjUITh5sD9kwsAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [FruitName = _t, Month = _t, Value = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"FruitName", type text}, {"Month", type date}, {"Value", Int64.Type}})
in
    #"Changed Type"

Rankings table:

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlSK1YlWMgKTxmDSBEyaKsXGAgA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Rank = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Rank", Int64.Type}})
in
    #"Changed Type"

[Tevon713]

@Greg_DecklerThank you very much, this will work for the time being.

[CNENFRNL]

Just enlighted by some PQ guru the other day, I came up with a verbose method to calculate rank instead of sorting,

Rank = List.Count( List.PositionOf( list, value, Occurrence.All, each _>=value ) )

 Just for fun, here's a PQ solution based on your hearty FruitBasket. Cheers!

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WciwoyElV0lHySszTNTIEMkwNDJRidRASbqlJEAkTNAnfxCKIhBFUIiA1sQjZJBNUcbhBxqjicHNgNjsl5gEhsknG6DJwszD0wE0zhMr4FyXmpacWIxtnhCEFNw9TCm4gzDvemeWZyKYZoorDjUITh5sD9kwsAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [FruitName = _t, Month = _t, Value = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"FruitName", type text}, {"Month", type date}, {"Value", Int64.Type}}),
    #"Added Custom" = Table.AddColumn(#"Changed Type", "YYYYMM", each Date.ToText([Month], "MMM yyyy")),
    #"Grouped Rows" = Table.Group(#"Added Custom", {"YYYYMM"}, {{"ar", each _, type table [FruitName=nullable text, Month=nullable date, Value=nullable number, YYYYMM=text]}}),
    #"Added Rank" = Table.TransformColumns(#"Grouped Rows", {"ar", (t) => Table.AddColumn(t, "Rank", (r) => List.Count(List.PositionOf(t[Value], r[Value], Occurrence.All, each _ >= r[Value])))}),
    #"Expanded ar" = Table.ExpandTableColumn(#"Added Rank", "ar", {"FruitName", "Rank"}, {"FruitName", "Rank"}),
    #"Pivoted Column" = Table.Pivot(#"Expanded ar", List.Distinct(#"Expanded ar"[YYYYMM]), "YYYYMM", "FruitName")
in
    #"Pivoted Column"