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

Compete to become Power BI Data Viz World Champion! First round ends August 18th. Get started.

Reply
Giuseppe_Aino
Advocate II
Advocate II

Stacked column chart unexpected disappearance of some histogram bars when higher than 99

It's strange behavior that can be explained by layout reasons but which is obviously misleading and which may lead one to believe in a lack of data that in reality is not there.

 

The assumptions verified are that:

  • the data for the days that are not represented with rectangles are real in table
  • the rectangles start to disappear when these become greater than 99
  • the rectangles that disappear are in symmetrical positions with respect to the center of the x-axis

Workaround (but which I cannot accept):

  • change the representation on the x-axis from "Continuous" to "Categorical"
  • change the visual to: line chart, clustered column chart
  • set the visualization to "Ribbons" (which is an illusion going to fill the empty spaces).

Working view (99 days):

Screenshot 01.png

For 100 days rectangle for March 23 disappears:

Screenshot 02.png

And so on, always worse:

Screenshot 03.png

Same graph but with lines, note that the data has always been there!

Screenshot 04.png

 

Please tell me if this issue it's a bug or intended behavior, thanks anyway and in advance

1 ACCEPTED SOLUTION

With this additional info I lean very heavily towards it being a rendering issue (i.e. you are trying to fit too many data points into the stacked column chart). 

 

Why you see it work on the line chart and not the stacked column chart is due to how rendering engines treat point-based vs. area-based visual encodings, especially with respect to the x-axis layout and value interpolation.  Stacked column charts handle overlap by clipping bars. 

 

A line chart places a vertex at each x-y coordinate and simply draws a line between them. If the x-axis spacing is enough for at least one pixel, it renders all points.

 

Column charts need to compute bar width based on the number of points and axis range. If a bar is too thin (e.g., <1 px), it may be dropped or clipped. Line charts bypass this entirely.

 

Please mark this post as a solution if it helps you. Appreciate Kudos.

View solution in original post

4 REPLIES 4
andrewsommer
Super User
Super User

I would bet this is a rendering limitation.  The fact that the disappearing bars occur symmetrically and only under specific conditions strongly suggests a layout engine or axis scaling issue, rather than a problem in the underlying data.  Just to make sure though, is the date field on the x-axis properly defined as a date field?

 

When the x-axis is set to Continuous, the chart rendering engine interprets it as a time or numeric scale with evenly spaced intervals, and applies internal logic to avoid "overplotting" or visual clutter.

 

When you switch the x-axis to Categorical, each data point is treated independently without interpolation. That means; each x-value is rendered, regardless of spacing and no auto-binning or interpolation occurs.

 

 

Please mark this post as a solution if it helps you. Appreciate Kudos.

Hi Andrewsommer, first of all thank you for your answer.

Your question about the date field on the x-axis struck me and I checked as follows.

On the x-axis I use the "DayStart" field which on the database is DateTime but which I then formatted on PowerBI as type date.

I don't know if it's important but then I use this field as a relationship with a calendar table to filter data on the page with the Slicer.

Reply Screenshot 01.png

Database side:

Reply Screenshot 02.png

It all seems coherent and aligned, the idea that I have to keep this strange behavior is gaining ground more and more.

Thank you

With this additional info I lean very heavily towards it being a rendering issue (i.e. you are trying to fit too many data points into the stacked column chart). 

 

Why you see it work on the line chart and not the stacked column chart is due to how rendering engines treat point-based vs. area-based visual encodings, especially with respect to the x-axis layout and value interpolation.  Stacked column charts handle overlap by clipping bars. 

 

A line chart places a vertex at each x-y coordinate and simply draws a line between them. If the x-axis spacing is enough for at least one pixel, it renders all points.

 

Column charts need to compute bar width based on the number of points and axis range. If a bar is too thin (e.g., <1 px), it may be dropped or clipped. Line charts bypass this entirely.

 

Please mark this post as a solution if it helps you. Appreciate Kudos.

This is true. If OP doesnt want such, then categorial must be selected. This or that.





Dane Belarmino | Microsoft MVP | Proud to be a Super User!

Did I answer your question? Mark my post as a solution!


"Tell me and I’ll forget; show me and I may remember; involve me and I’ll understand."
Need Power BI consultation, get in touch with me on LinkedIn or hire me on UpWork.
Learn with me on YouTube @DAXJutsu or follow my page on Facebook @DAXJutsuPBI.

Helpful resources

Announcements
August Power BI Update Carousel

Power BI Monthly Update - August 2025

Check out the August 2025 Power BI update to learn about new features.

August 2025 community update carousel

Fabric Community Update - August 2025

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

Top Solution Authors