Spark - Not all data processes in JavaRDD complex object - list

Question

I have the following code that reads in a text file of 5 rows CSV floats:
0.014, 0.035, 0.030, 0.018, 0.023, 0.027, 0.035, 0.036, -0.009, -0.013, 0.026, 0.042
0.032, 0.055, -0.036, 0.052, 0.047, 0.034, 0.063, 0.048, 0.025, 0.040, 0.036, -0.017
0.054, 0.056, 0.048, -0.007, 0.053, 0.036, 0.017, 0.047, 0.019, 0.017, 0.040, 0.032
0.038, 0.062, -0.037, 0.050, 0.065, -0.043, 0.062, 0.034, 0.035, 0.056, 0.057, 0.025
0.049, 0.067, -0.039, 0.051, 0.049, 0.037, 0.055, 0.025, 0.052, 0.020, 0.045, 0.040
The code loads in the data using Spark's JavaSparkContext textFile().
JavaSparkContext sc = new JavaSparkContext(master, "basicportopt", System.getenv("SPARK_HOME"), System.getenv("JARS"));
JavaRDD<String> lines = sc.textFile(".../src/main/Resources/portfolio.txt");
Next the data is loaded into a JavaRDD type as a List of Lists of type Double:
JavaRDD<List<List<Double>>> inputData = lines.map(new Function<String, List<List<Double>>>() {
#Override
public List<List<Double>> call(String s) {
List<List<Double>> dd = new ArrayList<List<Double>>();
double d = 0;
List<Double> myDoubles = new ArrayList<Double>();
for (String value : s.split(",\\s*")) {
d = Double.parseDouble(value);
myDoubles.add(d);
}
dd.add(myDoubles);
return dd;
}
});
Finally, the idea is that the data will be manipulated to produce some calculations to produce some summary results in the following algorithm:
inputData.foreach(new VoidFunction<List<List<Double>>>() {
public void call(List<List<Double>> col) {
System.out.println("Starting with first row...");
ArrayList l = (ArrayList) col.get(0);
for (List<Double> m : col) {
Double sum1 = 0.0;
for (Double d : m) {
sum1 += d;
}
Double avg1 = sum1 / m.size();
System.out.println("The avg of the row \"m\" being worked with: " + avg1);
System.out.println("Crunch the first fow with the other rows including self.");
for (List<Double> n : col) {
Double sum2 = 0.0;
for (Double d : n) {
sum2 += d;
}
Double avg2 = sum1 / m.size();
System.out.println("The avg of the row \"n\" being worked with: " + avg2);
Double xy = 0.0;
for (int index = 0; index < m.size(); index++) {
xy += m.get(index) * n.get(index);
}
xy -= (avg1 * avg2);
System.out.println("Resulting covariant: " + xy);
}
}
}
});
However I would expect to get 25 results I only get 5 results, because in the line:
for (List<Double> m : col) {...}
I would expect "col" to have 5 elements but stepping through the debugger shows only 1 element.
But using the collect() method:
List<List<List<Double>>> cols = inputData.collect();
shows 5 elements.
Why does the foreach() method not contain the 5 elements?