Power BI Limitations and Known Issues
This section describes limitations and known issues of AtScale's support for the Power BI SQL Server Analysis Services Connector.
Microsoft Limitations
Client-Side Modeling and Custom DAX Expressions
Microsoft Power BI does not support client-side modeling features or
custom DAX expressions for Live Connections to Multi-Dimensional data
sources such as Microsoft SQL Server Analysis Services and AtScale. To
create report-scoped calculations, you must either connect to AtScale
from Power BI in Tabular mode, or configure the AtScale engine to run in
tabular
mode. For more information, see Configuring
Connections.
Client-Side Multi-Datasource Modeling
Microsoft Power BI does not support multi-data source modeling with Live Connections to Multi-Dimensional data sources such as Microsoft SQL Server Analysis Services and AtScale. If you are interested in using client-side multi-datasource modeling with AtScale 2022.3.0 or later, ask your Account Executive about participating in AtScale's Tabular DAX Beta program.
Access to Cube-Side Drill-Through Sets
Microsoft Power BI does not support access to cube-side multi-dimensional Drill-through Sets, known as 'Actions' in SSAS, with Live Connections to Multi-Dimensional data sources such as Microsoft SQL Server Analysis Services and AtScale. There are no AtScale enhancements scheduled for this issue because it is a Microsoft Power BI limitation.
Decomposition Tree visualization
With respect to connections, Microsoft classifies AtScale as an 'On-premise SQL Server Analysis Services' data source. Microsoft Power BI does not support the Decomposition Tree visualization from such sources. For details, see Known Limitations.
Key Influencers visualization
With respect to connections, Microsoft classifies AtScale as an 'On-premise SQL Server Analysis Services' data source. Microsoft Power BI does not support the Key Influencers visualization from such sources. For details, see Considerations and Troubleshooting.
For more information regarding Power BI's SSAS MD data source connection capabilities and limitations, see Connect to SSAS multidimensional models in Power BI Desktop.
Queries with large numbers of chained ISBLANK constraints
When using the Legacy DAX dialect, Power BI generates queries with large numbers of non-user specified ISBLANK function calls, which can slow down query processing. If a cube is affected by this issue, you can connect to it using either the SuperDAX MD or DAX Tabular dialects; for more information, see Using SuperDAX MD and Configuring Connections.
Alternatively, you can use the following workaround:
- Make sure you are an Organisation Administrator or Super User.
- Access the cube settings; for details, see Cube Level Configuration Settings.
- Select the cube affected by the issue.
- Locate the
query.language.dax.bypassIsBlank
setting and set it to True. - Save your changes.
Note: When this setting is enabled, the cube will not be able to service intentional ISBLANK filters.
Known AtScale Issues
Measure Filtering
AtScale supports Power BI filters on Measure. Consider the following:
- To use this feature, set the Engine setting
QUERY.FACTLESS.IGNOREINCIDENTALFILTER = FALSE
and restart the Engine. - Power BI Measure filtering is NOT supported on AtScale Calculated Measures that use dimension modification functions, such as lag, parallelperiod, periodstodate, parent, siblings, children, descendants, ancestor, and so on.
- To allow dimensionally-modified measure filter queries, you can use
the
query.language.dax.blockDimModifiedMeasureFilters
engine setting. For details, see Changes related to upgrades.
Scatter chart is not displayed when using SuperDAX MD
When the connection from Power BI to AtScale is configured to use the SuperDAX MD dialect, trying to use a Scatter chart in Power BI leads to "Can't display the visual" error. (ATSCALE-9570)
Matrix visualization is not supported for CDP Impala 3.x
Matrix Visualization in Power BI is not supported for CDP Impala version 3.x. (ATSCALE-9552)
Date Filtering
Trying to apply a filter based on a columns of type Date could lead to issues with the visualization. As a work-around, you can convert the type from Date to DateTime or Timestamp using a calculated column, and then adjust the attribute column binding.(ATSCALE-6703)