Splitting List into sublists based on unique values

Question

I have a list of lists: List<ArrayList<String>> D = new ArrayList<>();

When it's populated, it might look like:
["A", "B", "Y"]
["C", "D", "Y"]
["A", "D", "N"]

I want to split the list of lists into partitions based on the unique attribute values (let's say index 1).

So the attribute at index 1 has two unique values, "B" and "D", so I want to split into:
["A", "B", "Y"]

["C", "D", "Y"]
["A", "D", "N"]

and put those into a List<ArrayList<ArrayList<String>>> sublists;

Is there a smart way of doing this, or do I just do something like this:

List<ArrayList<ArrayList<String>>> sublists = new ArrayList<>();
int featIdx = 1;

// generate the subsets
for (ArrayList<String> record : D) {
    String val = record.get(featIdx);

    // check if the value exists in sublists
    boolean found = false;
    for (ArrayList<ArrayList<String>> entry : sublists) {
        if (entry.get(0).get(featIdx).equals(val)) {
            entry.add(record);
            found = true;
            break;
        }
    }

    if (!found) {
        sublists.add(new ArrayList<>());
        sublists.get(sublists.size()-1).add(record);
    }
}

This is a step from the C4.5 Decision Tree algorithm, so if anyone has experience in this, I would appreciate if you could let me know if this is the right approach to generating the sublists.

Thank you.

Answer 1

With Java 8 you can use the groupingBy collector:

Map<String, List<List<String>>> grouped = D.stream()
                .collect(Collectors.groupingBy(list -> list.get(1)));
Collection<List<List<String>>> sublists = grouped.values();

or as suggested by @AlexisC:

import static java.util.stream.Collectors.collectingAndThen;
import static java.util.stream.Collectors.groupingBy;

Collection<List<List<String>>> sublists = D.stream()
             .collect(collectingAndThen(groupingBy(list -> list.get(1)), Map::values));