Code for demystifying spark serization error article

spark-error.scala

/**
 * Model for our case.
 * Note that [[Value]] class is not serializable, while
 * [[Wrapper]] is a case class, which is serializable by default.
 */
class Value
case class Wrapper(v: Value)

// defined class Value
// defined class Wrapper

val wrapper = Wrapper(new Value)
// wrapper: Wrapper = Wrapper(Value@2d8f65a4)

val data = sc.parallelize(1 to 10)
// data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[1]

data.map(_ => value).foreach(println)
// org.apache.spark.SparkException: Task not serializable
// at org.apache.spark.util.ClosureCleaner$.ensureSerializable(ClosureCleaner.scala:304)
// ...

sc.parallelize(1 to 10).map(_ => Wrapper(new Value)).foreach(println)
// Wrapper($line15.$read$$iwC$$iwC$Value@1479edee)
// Wrapper($line15.$read$$iwC$$iwC$Value@342efa0e)
// 8 more ...

data.map(_ => Wrapper(new Value)).collect()
// 15/11/16 10:07:38 ERROR Executor: Exception in task 6.0 in stage 1.0 (TID 14)
// java.io.NotSerializableException: $line15.$read$$iwC$$iwC$Value
// ...

data.map(_ => Wrapper(new Value)).count()
// res6: Long = 10

data.map(_ => Wrapper(new Value)).map(_.toString).reduce(_ + _)
// res7: String = Wrapper(...)Wrapper(..)...