Hi, Just wondering if anyone can shine some light on the idea. I would like to write an application to listen to a hive table (which is slowly adding partition when data comes in) then write the newly arrived data to kafka. The reason why the upstream job (which writes to the hive table) can’t directly write to kafka is complicated, so I want to leave it out of the discussion. Can flink do that? I might use the wrong keyword to do a search because I found nothing, is there anyone did similar thing and can share with me? Thank you so much.