Hadoop Stores data in local disk Slow speed Suitable for batch processing External schedulers required High latency No in-built interactive mode. less expensive hardware very difficult to work. Spark Stores data in-memory Faster speed Suitable for batch and real-time processing Schedules tasks itself Low latency Has interactive mode Lot of RAM to run in-memory, increasing it in the cluster, gradually increases its cost. It is easy to program ... View more

Just before the command use the ! mark, we run unix shell commands from the hive. For example- ! pwd at hive prompt will list the current directory. ... View more