O'Reilly logo
  • HemantKumar Thakur thinks this is interesting:

sc.parallelize(data).reduceByKey((x, y) => x + y) // Custom parallelism

From

Cover of Learning Spark

Note

Missing parameter in custom partitioning