MongoDB Connector for Spark V10 and Change Stream

I read mongodb change stream and write to kafka. No matter how I change read.change.stream.publish.full.document.only it doesn’t work.

I want to get metadata (operationType etc) but the result only show fullDocument.