Highest value per cluster

[robertosangi]

Hi, 

In my case I want to obtain the highest value (top 1) of a column divided per cluster. Here the example:

Actually I filtered my column as per picture because is not possible to filter per TOP N:

As you can se from the image "Palermo Ovest" in this case has only one row and it is ok but "Partinico" has different values because of many cases greater than 10. 

How can I show only the highest per each cluster in bold.

Thanks 

[amitchandak]

@robertosangi , Create a rank on the innermost column and filter on 10

rankx(allselected(Table[cluster]), [IGB],,desc,dense)

[robertosangi]

Hi @amitchandak , i don't want to filter on 10 but on the maximum value. It is possible?

[Anonymous]

Hi @robertosangi ,

Do you want to filter your matrix visual by two filters, 1. values in IGB_Noloc are greater than 10, 2. maxest value in  IGB_Noloc in per cluster ?

Here I suggest you to create a measure to filter your visual.

My sample:

Measure:

IGB_NoLoc Greater than 10 and Top 1 =
VAR _IGB_NoLoc =
    CALCULATE (
        SUM ( 'Table'[Value] ),
        FILTER ( 'Table', 'Table'[Type] = "IGB_NoLoc" )
    )
VAR _MAX_IGB_NoLoc =
    CALCULATE (
        MAX ( 'Table'[Value] ),
        FILTER (
            ALL ( 'Table' ),
            'Table'[Unite] = MAX ( 'Table'[Unite] )
                && 'Table'[Level2] = MAX ( 'Table'[Level2] )
                && 'Table'[Type] = "IGB_NoLoc"
                && 'Table'[Value] > 10
        )
    )
RETURN
    IF ( _IGB_NoLoc = _MAX_IGB_NoLoc, 1, 0 )

Add this measure into visual level filter of matrix and set it show items when value =1.

Result is as below.

Best Regards,
Rico Zhou

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[robertosangi]

Hi @Anonymous thanks for your reply. I want to filter only per the highest value if possibile. Not filtering per 10 value

[Anonymous]

Hi @robertosangi ,

I think my measure will show you the highest value in IGB_Noloc. If this still couldn't help you solve your problem, please show me a screenshot with the result you want, this will make me easier to find the solution.

Best Regards,
Rico Zhou

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[robertosangi]

Hi @Anonymous,

Sure, I'll try to explain better

As you can see this is the AS IS situation. I want, per each red category, the highest IGB_NoLoc value. So, in this case for PALERMO EST only the first and the third value. In PALERMO OVEST only the first.

It is clear? I'll try to explain better if not.

Thanks

[Anonymous · Accepted solution]

Hi @robertosangi ,

According to your screenshot, I think your table should look like as below.

I think you want to show values in red box.

I suggest you to create a rank measure and use it as a filter in visual level filter.

RANK =
VAR _RANK =
    RANKX (
        ALLEXCEPT ( 'Table', 'Table'[Unite], 'Table'[Blue Team] ),
        CALCULATE ( SUM ( 'Table'[IGB_NoLoc] ) ),
        ,
        DESC,
        DENSE
    )
RETURN
    _RANK

Result is as below.

Best Regards,
Rico Zhou

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.