I'm running spark 3.0.2, with spark-cassandra connector assembly 3.0.0 on spark - cassandra cluster as each node is a spark worker and have cassandra instant running on same machine.
val xRDD= X.repartitionByCassandraReplica(Keyspace,Table)
that xRDD result became blank.
does this still be supported in 3.0.2, or there is any reason could let the rdd after using the repartitionByCassandraReplica empty ?
is it really affect the performance if I called the "joinWithCassandraTable" without doing the repartitionByCassandraReplica ?
exception in the log :
Could not load org.apache.tinkerpop.gremlin.process.traversal.dsl.graph.GraphTraversal with loader null: java.lang.ClassNotFoundException