org.apache.spark.sql.execution.datasources.json
Create a new RDD via the supplied callback if there is at least one file to process, otherwise an org.apache.spark.rdd.EmptyRDD will be returned.