@Michael_Appiah circling back on this because the forEachBatch functionality has landed in Lakeflow:

https://docs.databricks.com/aws/en/ldp/for-each-batch

So now you can do this in "DLT". But there's going to be a lot of boiler plate stuff like error handling, full refresh cleanup, logging that needs to be done manually compared to the normal functonalities.

@adriennn this has nothing to do with DLT, but about Databricks providing a different session implementation here than regular Spark.

@Malthe  the question did not indeed pertain to DLT / lakeflow, but another user asked about it and I mistakenly mentionned OP instead of @alex0sp  
cheers!