Adding Legend to Stacked Bar graph distorting the %s

[F_Reh]

Good Afternoon,

I have following Stacked Bar graph (Diag 1):

When I add the Categories Band field to the legend (in order to show the break down within each bar) there seems to be a problem (Diag 2):

For example, for November 2025 when I calculate the %s for all categories it is indeed 3.9% (as shown in Diag 1), but when I simply add the Categories Band field to the Legend then each segment (which appears) is around 4-5%, and the total bar extends to about 23% (rather than 3.9%) !

Kindly advise

[v-pnaroju-msft]

Hi F_Reh,

Thankyou for the followup.

The expression SELECTEDVALUE(Parameter_HI[ParameterLabel], "Organisation Action/Client Action #") uses the string default so the measure has a predictable fallback when the slicer is empty or multi-selected. In the PBIX file, Parameter_HI is a disconnected table created using Enter Data, not the built-in Field Parameter. The disconnected approach is straightforward and works with the SWITCH wrapper to map the slicer label to the appropriate # or % measure.

Additionally, please find attached a screenshot and a revised sample PBIX file that uses the Field Parameters feature, which may assist in resolving the matter:

We hope the information provided will help resolve the issue. Should you have any further queries, please feel free to contact the Microsoft Fabric community.

Thank you.

[F_Reh]

Thank you...In relation to your measure, why is "Organisation Action/Client Action #" passed in to SELECTEDVALUE(), as well in inside the RETURN statement ?

Selected_Measure_Value =

VAR sel = SELECTEDVALUE( Parameter_HI[ParameterLabel], "Organisation Action/Client Action #" )

RETURN

SWITCH(

  TRUE(),

  sel = "Organisation Action/Client Action #", [OrgClient_Count_Measure],

  sel = "Organisation Action/Client Action %", [OrgClient_Pct_of_Month],

  sel = "Organisation Action #", [Org_Count_Measure],

  sel = "Organisation Action %", [Org_Pct_of_Month],

  sel = "Client Action #", [Client_Count_Measure],

  sel = "Client Action %", [Client_Pct_of_Month],

  BLANK()

)

In your PBIX file is Parameter_HI a Field Parameter, or is it a simple Table ?

Am confused by the measure in your file, because I only made Field Parameters in the past by very simply inserting columns & measures:

Regards

[v-pnaroju-msft]

Thankyou, @rohit1991 and @anilelmastasi for your response.

Hi F_Reh,

We appreciate your question on the Microsoft Fabric Community Forum.

Based on my understanding of the scenario, please find attached a screenshot and a sample PBIX file that may assist in resolving the matter.

We created a measure that ignores the legend filter using MonthDenominator. Additionally, we created a single wrapper measure, Selected_Measure_Value, and used this measure in the visual.

We hope the information provided helps to resolve the issue. Should you have any further queries, please feel free to contact the Microsoft Fabric community.

Thank you.

[F_Reh]

Sorry, if I am not mistaken but in previous post I can't see references to any Columns/Measures (other than the one passed on to the REMOVEFILTERS function) so I haven't tried that one as of yet.

Anyway, I looked at the earlier recommended solutions and tried this:

Organisation Action/Client Action %" = CALCULATE( DIVIDE(SUM(<Table>[Numerator],<Table>[Total/Denominator]),REMOVEFILTERS(<Table> [Age_Band]))

However, it still produces relatively huge %s for each segment:

But when I remove the Age field from the Legend, the Total Monthly %s  (for all Age bands) become exactly right (checking over with manual calculation)

[F_Reh]

Thanks for the great explanation. My problem is complicated by having a slicer with six options pertaining to a business scenario relating to whose responsibility it was for a specific action (whether Organisation or Client, or both). The six options are linked to a Field Parameter:

Parameter_HI = {

    ("Organisation Action/Client Action #", NAMEOF(‘Table'[Organisation Action/Client Action (KPI) #]), 0),

    ("Organisation Action /Client Action %", NAMEOF(‘Table'[Organisation Action/Client Action (KPI) %]), 1),

    ("Organisation Action #", NAMEOF(‘Table'[Total Organisation Action (KPI) #]), 2),

    ("Organisation Action %", NAMEOF(‘Table'[Organisation Action (KPI) %]), 3),

    ("Client Action #", NAMEOF(‘Table'[Total Client Action (KPI) #]), 4),

    ("Client Action %", NAMEOF(‘Table'[Client Action (KPI) %]), 5)

}

The visuals on the Page separately display the Diversity bands (Age, Ethnicity, Gender etc).....

In light of using Field parameter based Slicer, each option with their respective allocated measure will apply that measure to all the Diversity visuals (on the page).....

So I feel it will be extremely cumbersome to create new measures for each of the Bands for each of the Diversity Visuals, and also taking in to account that each option on the Slicer applying to the whole page will inherently only be linked with one Measure at a time. 

[anilelmastasi]

@F_Reh ,

You can use SELECTEDMEASURE() Function. Like this:

Universal % of Total Month =
VAR CurrentMeasure = SELECTEDMEASURE()
RETURN
DIVIDE(
CurrentMeasure,
CALCULATE(
CurrentMeasure,
REMOVEFILTERS(CategoryBand
)

[F_Reh]

Thanks. If I can ask for clarification...My current measure is simply as follows:

Measure % = DIVIDE(SUM(<Table>[Column X], <Table>[Column Y])

I require <Table>[Column Y] in the denominator to calculate the percentage which X represents of the entire dataset.

In the example given - from what I am seeing -  SUM(num) is in both numerator and denominator, and also the REMOVEFILTERS seems only applied to the denominator ?

[anilelmastasi]

When Categories Band is added to Legend, Power BI applies a filter:

Like:

Category Band = A

Category Band = B

Category Band = C

So the measure becomes:

For Category A: SUM X (Category A) / SUM Y (Category A)

For Category B: SUM X (Category B) / SUM Y (Category B)

This is not a % of the whole dataset anymore.It is % of each category, so each percentage is much larger.

You want X for that Category Band / Y for the entire dataset (not filtered by Category Band) To achieve this, you must remove the legend filter from the denominator only using REMOVEFILTERS.

If this solved your issue, please mark it as the accepted solution. ✅

[anilelmastasi]

Hello @F_Reh ,

Once you add Category Band to the Legend, Power BI calculates the measure at Month + Category Band granularity.

% by Category of Total Month =

DIVIDE( SUM(num),

CALCULATE(SUM(num),

REMOVEFILTERS(CategoryBand)) )

Simply  you can use this DAX.

If this solved your issue, please mark it as the accepted solution. ✅

[rohit1991]

Hii @F_Reh 

If you want to keep the 3.9% total while also showing the category split, you must create a measure that ignores the legend filter using:

% All Categories = CALCULATE([Your % Measure], REMOVEFILTERS('Category Band'))

Then use this measure for the bar length and keep the legend for the color split.