Forum Discussion
Dataverse table access fails with TDS error 8657 in Fabric Dataflow
The issue has been resolved with a workaround. I split the source query into two smaller queries and then joined them back together, keeping the existing final table structure unchanged. This avoids the TDS connector error caused by the large number of columns in the project table.
The data is now refreshing successfully.
Microsoft still needs to investigate and resolve the underlying Dataverse connector/TDS issue because I am still unable to preview the data through the normal Dataverse connector.
Because other tables work and a new Dataflow reproduces the failure before transformations, investigate the msdyn_project source query separately from the Lakehouse or semantic model. Connect to the same Dataverse TDS endpoint in SQL Server Management Studio using the same authorized identity. Test a TOP 10 query selecting only the table's primary-key column, then add required columns in small groups. Microsoft recommends limiting rows and columns; choice labels and large text columns can make these queries more expensive. If the minimal query succeeds, compare it with the Dataflow query and test disabling relationship columns through CreateNavigationProperties=false in a separate test query. That option addresses documented metadata overhead, not a confirmed fix for error 8657. If even the minimal query fails, retain its exact text, UTC timestamp, request identifier, and error details for Dataverse support. Do not change production metadata based on this error alone.
Hi Jamony,
Thank you for your response, I am not able to preview the data at all, even in a brand-new Dataflow with no transformations. Since other Dataverse tables work and this failure occurs before transformations, Lakehouse loading, or the semantic model, this points to a Dataverse connector/TDS issue.