Index depending on column value

[robertosangi]

Hi,

I have this situation:

I created a conditional column that indexed my value. My problem is that I want this count restart depends on the week value and more i want to index my values depends on "Count" column. 

So i want to create a sort of rank of the week per each value of the "Codice Unità Territoriale" column.

How I can do that?

Thank you

[serpiva64]

E' stato un piacere!

[serpiva64]

Hi,

From this:

you canobtain this

applying these steps:

let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("Vc45DoAwDETRu7hOwUSsZ4ko2EFiuX8H2E4xRYqn2NZPSQYJEr9XSh+SjK5WNbmicXY2qoUW1zzaKTenac+fhfLgs2dmrbzo7u2CzT7OSvW3g9pB7eB2UDuoHdwOage3g9vB7aB2cDusvX8B", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [CoUnTerr = _t, Week = _t, Count = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"CoUnTerr", type text}, {"Week", Int64.Type}, {"Count", Int64.Type}}),
#"Grouped Rows1" = Table.Group(#"Changed Type", {"Week", "Count"}, {{"AllRows", each _, type table [CoUnTerr=nullable text, Week=nullable number, Count=nullable number]}}),
#"Grouped Rows" = Table.Group(#"Grouped Rows1", {"Week"}, {{"AllRows", each _, type table [CoUnTerr=nullable text, Week=nullable number, Count=nullable number]}}),
#"Added Custom1" = Table.AddColumn(#"Grouped Rows", "Custom.1", each Table.Sort([AllRows],{{"Count", Order.Descending}})),
#"Added Custom" = Table.AddColumn(#"Added Custom1", "Custom", each Table.AddIndexColumn([Custom.1],"Index",1)),
#"Expanded Custom" = Table.ExpandTableColumn(#"Added Custom", "Custom", {"AllRows", "Index"}, {"AllRows.1", "Index"}),
#"Expanded AllRows.1" = Table.ExpandTableColumn(#"Expanded Custom", "AllRows.1", {"Count"}, {"Count.1"}),
#"Removed Columns" = Table.RemoveColumns(#"Expanded AllRows.1",{"AllRows", "Custom.1"})
in
#"Removed Columns"

If this post is useful to help you to solve your issue consider giving the post a thumbs up 

 and accepting it as a solution !

[robertosangi]

Hi @serpiva64 

Could you please write the steps without code? Is not well understandable but seems work.

Thanks

[serpiva64 · Accepted solution]

From your table

you group rows by week and by count (in order to manage situation of parity)

hten you group by week

then you order descending your column count inside your groups

at this point you add and index column

at this point you first expand Custom

you obtain this

Finally you expand AllRows.1 

remove unnecessary column and that's done

[robertosangi]

Hi  @serpiva64 I reached this step:

but expanding i obtained:

I think I missed something but i'm reaching the goal

[serpiva64]

Sorry i forgot to expand one column. You can do it in

and this is the final result

[robertosangi]

@serpiva64 just to close the question. 

I want to represent this in table. I wanted to insert the data you found in such way:

How can I do that in a table?
Thanks

[serpiva64]

Hi,

You can achieve something like this from your previous 

Add acolumn with W-1

Merge your query on itself

Expand

Change type then add difference column

change type

then you can create your matrix

[amitchandak]

@robertosangi , refer if this video from Curbal can help

https://www.youtube.com/watch?v=7CqXdSEN2k4

[robertosangi]

Hi @amitchandak 

I watched the video, is similar to my case but doesn't work If I follow the steps she shown. 

Actually is not preciselly my case and for this reason gives me several errors