Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
Description
When building a cube from Kafka topic, the first step is to sink the Kafka data to HDFS. In today's implementation, it will persist all the fields of a message to disk. While in many cases, only a couple of fields will be needed for cubing; Today's behavior wastes network bandwidth and disk space.