Collectors: interview questions & scenarios

Model answer

A collector is a recipe for turning many stream elements into one result. The five pieces are supplier, accumulator, combiner, finisher and characteristics.

The supplier creates a fresh mutable accumulation container. The accumulator folds one element into that container. The combiner merges two partial containers, which is what parallel streams need. The finisher converts the work container into the published result, unless the collector has IDENTITY_FINISH.

Collector<String, StringJoiner, String> csv = Collector.of(
        () -> new StringJoiner(","),
        StringJoiner::add,
        StringJoiner::merge,
        StringJoiner::toString
);

The combiner matters because a sequential stream might never expose a bad one. A collector with a wrong combiner can pass local tests and then drop data, duplicate data or reorder data when somebody changes stream() to parallelStream().

Model answer

import java.util.*;
import java.util.stream.*;

public class Main {
    record User(String id, String name) { }

    public static void main(String[] args) {
        List<User> users = List.of(
                new User("u1", "Mina"),
                new User("u2", "Ravi"),
                new User("u1", "Mira")
        );

        System.out.println(users.stream().collect(
                Collectors.toMap(User::id, User::name, (oldName, newName) -> oldName)
        ));

        users.stream().collect(Collectors.toMap(User::id, User::name));
    }
}

The first collection succeeds because it gives a duplicate-key policy: keep the old value. A typical Java 21 output is:

{u1=Mina, u2=Ravi}
Exception in thread "main" java.lang.IllegalStateException: Duplicate key u1 (attempted merging values Mina and Mira)

Do not rely on the HashMap print order as the point of the question. The important part is that the two-argument toMap overload treats duplicate keys as a bug and throws IllegalStateException.