Microsoft has introduced a new capability in Power BI Desktop that allows users to download big semantic models as.pbix files. Previously, altering reports associated with these models was only allowed in live connected report mode. Users may now open and edit both reports and large models as.pbix files directly in Power BI Desktop. This improvement enables the complete end-to-end Desktop authoring experience for huge semantic models.
Large semantic models in Power BI Premium
Power BI semantic models may store data in a highly compressed in-memory cache, resulting in speedier queries and a better user experience. The huge Semantic Model Storage Format feature in Premium capacity allows you to enable huge semantic models that exceed the default size restriction. When enabled, the model size is restricted only by the Premium capacity size or the maximum size specified by the administrator.
Large semantic models can be enabled for all Premium P and Embedded A SKUs, as well as Premium Per User (PPU). The size restriction for these huge models in Premium is comparable to that of Azure Analysis Services.
To build semantic models greater than 10 GB, you must choose the Large Semantic Model Storage Format option. This setting offers other advantages, such as boosting the efficiency of XMLA write operations, even on smaller models. If you intend to use XMLA endpoint-based tools, enable this option.
Remember that huge semantic models in the service have no effect on the Power BI Desktop upload size restriction, which remains at 10 GB. However, models may exceed this limit during a service refresh.
Limitations when downloading a semantic model .pbix file
- To download the .pbix file, you must be at least a Contributor to the workspace.
- Semantic models with incremental refresh can't be downloaded to a .pbix file.
- Semantic models migrated from Azure Analysis Services by using the Azure Analysis Services to Power BI Premium migration feature can't be downloaded to a .pbix file. Downloading a .pbix file after deployment isn't supported as part of the deployment pipeline (ALM).
- Semantic models modified by using the XMLA endpoint can't be downloaded to a .pbix file.
- You can't download semantic models based on data imported from OneDrive for work or school.
- You can't download Direct Lake semantic models.
- Downloading a .pbix is subject to size limitations. The maximum sized allowed depends on the minimum of two factors: maximum offline semantic model size for the capacity SKU and the customizable max offline dataset size. Downloading a .pbix where the data model size is larger than 50 GB is not supported. Downloading a .pbix file that's larger than any of these limits is only supported when the .pbix is in live connected report mode.
- Semantic models with automatic aggregations enabled can't be downloaded to a .pbix file. You must disable automatic aggregations training in the semantic model settings and delete the aggregations tables by selecting the checkbox Delete automatic aggregations tables.
- Attempting to concurrently download a semantic model enabled for large semantic model storage format may result in errors.
- Semantic models enabled for large semantic model storage format cannot be downloaded using REST APIs. Use the Power BI Service to download these semantic models.