Hi Taranjeet Malik •,
Did you try below:
The recommended approach currently is to use Cosmos DB Spark Connector. This sample leverages change feed, bulk support, and structured streaming to read all historical data from beginning along with “deltas”:
Thanks