Microsoft Fabric Error:
Power BI Entity Not Found at Lakehouse Refresh
What does this error mean?
A Power BI semantic model connected to a Fabric Lakehouse (Direct Lake or import mode) failed to refresh because the underlying Lakehouse table referenced in the model was renamed, deleted, or is no longer accessible. The model's data mapping points to a table that no longer exists.
Common causes
- 1A Lakehouse table referenced by the semantic model was renamed or deleted since the model was last published
- 2A Spark notebook or pipeline dropped and recreated the table with a different schema, breaking the Direct Lake connection
- 3The semantic model was published before the Lakehouse table was created (table never existed during the first refresh)
- 4Table names have different casing after a schema change — Direct Lake is case-sensitive
How to fix it
- 1In the Fabric Lakehouse Explorer, verify that the table referenced in the model still exists with the same name and casing.
- 2If the table was renamed, update the semantic model's table mapping to point to the new table name and republish.
- 3If using Direct Lake mode, run a full refresh from the Fabric portal to revalidate all table references.
Beyond the docs
Common practitioner solutions not covered in the official documentation.
- 1Open the semantic model in Power BI Desktop or the Fabric web interface and check which tables are producing the error
- 2After schema changes to a Lakehouse table (add/drop columns), republish the semantic model to reflect the new structure