Spark batch reading from Kafka & using Kafka to keep track of offsets

Question

I understand that using Kafka\'s own offset tracking instead of other methods (like checkpointing) is problematic for streaming jobs.

However I just want to run a Spark b