1. Add SparkEtlJob class 2. Remove DppResult comment 3. Support loading from hive table directly #3433