@Rigobert @sat_db 

For schema evolution, the use of the model.json to build the spark schema should allow to cover all the supported cases of schema evolution that do not require a unlink/relink of the table, i.e. data colum type change.

If it is possible, you should also give a try to the Lakeflow Connect D365 Connector: https://learn.microsoft.com/en-gb/azure/databricks/ingestion/lakeflow-connect/d365