Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!

Reply
RoelVB
Helper I
Helper I

column chart showing different result than area chart

Hi,

 

I have this report that doesn't make any sense to me. My results in a stacked column chart are different that in a stacked line chart:

image.png

Note that the green area is missing in the column chart.

The area chart is just a copy of the column chart with the visual switched to the area chart.

 

The make things even stranger, if I deselect Q1 both charts are fine:

image.png

It doesn't have anything to do with Q1, because it's still fine if I select Q4 2019 till Q3 2020.

 

If I remove the legend from the column chart the totals is also fine:

image.png

 

A basic representation of the datamodel:

In the charts I'm using 3 columns from 2 tables.

As value the column "Amount" from "Projects".

As axis the column "Date" from "Projects".

As legend the column "Clientname" from "Clients".

The two tables have a one-to-many relation.

 

I really hope someone can see something that could be wrong here.

It looks like a bug in the bar chart to me, but if that's the case I would expect others to notice this.

 

Kind regards,

Roel

 

8 REPLIES 8
RoelVB
Helper I
Helper I

I did a lot of testing. And it gets crazier.

 

Here I use the ClientId as legend (this is NOT the related table, it's the ID inside the Projects table):

image.png

 

Here I use an anonymized client name, and everything is fine:

image.png

 

And with the actual client names:

image.png

I checked for double clientnames, but there are none.

 

Here's an anonymized PBIX: https://we.tl/t-8zS8CTHq8G

Hi @RoelVB ,

 

When I open the sample report you provided, this information is visually prompted, which will result in incomplete visual display.

V-lianl-msft_0-1612431067318.png

 

Unfortunately, there is no better way than to pre filter useless data in clientid field.

 

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

Thank you for checking my sample.

 

That's strange. I don't get this warning, but this still raises some new questions:

  1. Why the difference between using the client's ID vs Name as value? There are the same amount of names as IDs
  2. Why the difference between de column chart and area chart?
  3. I don't have anything to "pre filter", we need all data

Maybe it's just me, but this makes Power BI hugely unreliable.

The person who found this issue also didn't get a warning about "too many values" and uses this report to manage his department.

V-lianl-msft
Community Support
Community Support

Hi @RoelVB ,

 

Please try to use the date in "one" table as slicer. If the problem still exists, please recreate a pbix with similar model and remove sensitive data to share with us.

How-to-provide-sample-data-in-the-Power-BI-Forum 

 

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

Anonymous
Not applicable

Hello @RoelVB ,
It'll be more helpful if you provide some sample data or pbix file after removing the sensitive data.

Hi @Anonymous 

 

That's the problem. My datasource is SharePoint Online lists, I'm not sure how to remove sensitive data from the pbix file.

 

I need the clientnames to reproduce the issue, but I cannot publish a pbix including the clientnames.

Anonymous
Not applicable

Hello @RoelVB 
You may try to make a new visual instead of copying it.

First thing I tried, didn't help.

Helpful resources

Announcements
April AMA free

Microsoft Fabric AMA Livestream

Join us Tuesday, April 09, 9:00 – 10:00 AM PST for a live, expert-led Q&A session on all things Microsoft Fabric!

March Fabric Community Update

Fabric Community Update - March 2024

Find out what's new and trending in the Fabric Community.