RDD Exception Handling

[Kevin Conaway]

How can we catch exceptions that are thrown from custom RDDs or custom map
functions?

We have a custom RDD  that is throwing an exception that we would like to
catch but the exception that is thrown back to the caller is a
*org.apache.spark.SparkException* that does not contain any useful
information about the original exception.  The detail message is a string
representation of the original stack trace but its hard to do anything
useful with that.

Below is a small class that exhibits the issue.  It uses a map function
instead of a custom RDD but the symptom is the same, the original
*RuntimeException* is lost.  I tested this with spark 1.2.1 and 1.3.0


public class SparkErrorExample {

    public static void main(String [] args) throws Exception {
        SparkConf sparkConf = new
SparkConf().setAppName("SparkExample").setMaster("local[*]");
        JavaSparkContext ctx = new JavaSparkContext(sparkConf);

        JavaRDD<String> data = ctx.parallelize(Arrays.asList("1", "2",
"3"));

        try {
            data.map(line -> {
                throw new RuntimeException();
            }).count();
        } catch (Exception ex) {
            System.out.println("Exception class: " + ex.getClass());
            System.out.println("Exception message: " + ex.getMessage());
            System.out.println("Exception cause: "+ ex.getCause());
        }
    }
}