@sumitPanda  the solution builds the schema to read the CSV from the schema JSON file.

If the columns are added or removed (as per the Synapse Link itself), manual intervention is not needed. However, if a column type is changed, you need to drop the table and restart the ingestion (as you also have to unlink and link the table in the Synapse Link again).

This blog proposes an alternative solution to the Synapse Spark pools for those who are looking for a solution based on Databricks.

FYI, there is a Lakeflow Connect Connector for Dynamics. Just reach out to your Databricks account team for information.