What is one advantage of Spark’s in-memory data processing compared with MapReduce?
Spark can keep intermediate results in memory rather than repeatedly writing them to disk between processing stages. Reducing disk I/O can substantially speed up data processing.
Community Discussion