Common Chart Errors
Solutions and suggestions for common errors in charts.
Use the following steps to simplify your chart and identify the cause of the error:
Create a Table alongside the Chart using the exact same model, linked dimensions, and structure as the Chart; verify if the values are refreshed.
Explore the measures directly using the Data View; this reduces interference from filters and other specific chart features.
Copy the chart to a new canvas page and try to remove as many of the following as you can: measures, dimensions, and active filters.
Applying the Top N filter automatically is meant to help you avoid having too many overlapping data points in your charts while giving you enough data to allow you to adjust the filters or the structure of your charts.
The automatic Top N value is triggered at a different number of records for each chart type, but you can override the value if you manually set a Top N value that is larger than the default value.
For stacked chart types there is a stack or grouping limit that may be different from the automatic limit.
The following table shows the Top N trigger values (number of records) that are valid as of version 2020.02
|Chart type||Stack Limit||Total Records Limit||Mobile Total Records Limit||Scrollbar|
|Bar / Column||500||20||Yes|
|Bubble||No Limits Defined|
|Bullet||No Limits Defined|
|Combination Column & Line||500||20||Yes|
|Combination Stacked Column & Line||25||500||20||Yes|
|Scatterplot||No Limits Defined|
|Trend Time Series||No Limits Defined|
|Waterfall||No Limits Defined|
A time dimension exists and is bound to the outer-most dimension.
The chart supports Sort.
The chart has no existing sort or a sort exists on the time dimension.
If the auto limit is triggered, then it will automatically query and sort (ascending) the last-N values sorted by the Time dimension, provided that the Time dimension exists and is the outer-most dimension.
This message is likely due to a problem with the underlying models used in the story:
If you have multiple models, test one by one using the Data View option.
If you are using Live models, verify that the connection to the data source works as expected.
This message is displayed when the chart encounters an unexpected exception during rendering.
It is likely that you need to open an incident with Product Support.
To speed up troubleshooting, capture the following information using Chrome Developer tools (F12 or Ctrl + Shift + I ):
With Chrome Developer tools open, attempt to refresh the specific chart that fails.
Expand the error being thrown and capture the entire stack trace.
Attach these traces to the Product Support incident along with any data on the structure and formatting settings of the chart.
This behavior could be due to many different reasons, but it's possible that a component at a lower level has changed and now returns different data.
Use the following tips to find what might have changed:
Start with First steps for troubleshooting.
Identify if the artifact that is consumed has changed (Calculation view, CDS view, and so on).
Compare the data retrieved using other tools (if possible) such as HANA Design studio, Lumira, and so on.
The story was likely saved when the chart was in an error state.
Use the following tips to clear the error message:
Refresh the story.
Verify that all widgets are no longer in an error state.
Re-save the story.
When designing a chart, some of the labels may not appear when the chart is displayed.
The default setting for labels in charts is that when they overlap or collide with one another, or can't fit within the allotted space, they should be hidden. This ensures that the chart labels are readable and do not clutter the visualization unnecessarily. To see the value for a data-point, you can hover over it (or touch it on a touch display).
Causes of data label collisions
Labels in column charts are too long and span outside of the category, colliding with adjacent labels.
For overlapping data-points in a bar/column chart, only the largest absolute value of the overlapping labels will be shown.
This behavior can be modified by changing the Data Label styling options as explained in the help guide. See Styling Options for Charts.
Labels may be hitting or surpassing the bounds of the rendering area in the chart. In charts such as line and combined bar/column+line, the labels will attempt to be rendered in new positions close to the data-point. If that is not possible, then the label is hidden.
How to resolve data label collisions
Increase the size of the chart on the story.
Adjust axis label and data label font sizes.
Filter your data to allow more space.
Adjust the numeric scale and number of decimal places of the measure to higher magnitude that will shorten the label. For example, 15,000,000.00 unformatted could be scaled to 15 in millions with no decimal points.
Turn on the Round Data Labels option to allow decimal places in the data label to be rounded and conserve room before the label hides
If labels are overlapping, then the Avoid Data Label Overlap option has probably been unchecked in the Data Label styling options area for that chart. Re-enable that option to prevent labels from overlapping.
When you export a story to PDF, some of the charts in the PDF are blank.
When processing the export to PDF, each widget (chart, table, and so on) has a default timeout setting of three minutes. If it takes longer than three minutes to render the widget, the widget will be blank in the PDF.
You can ask your support person to enable a toggle for a longer rendering period (WIDGET_RENDER_COMPLETE_LONG_TIMEOUT), but be aware that longer timeout periods may affect system performance.