从 List 对象创建一个 SortedMap，其 Value 表示为映射到特定 Key 的 N 个最低对象属性的列表

Question

I am working with a CSV file which includes some information about accidents.

I've created the Accident type:

private Integer driverAge;
private Integer vehicleAge;

public Accident(Integer driverAge, Integer vehicleAge) {
    this.driverAge = driverAge;
    this.vehicleAge = vehicleAge;
}

I've also created a function that reads all the CSV file, converts all the accidents to a List<Accident> and saves it to this type AccidentArchive :

private List<Accident> accidents;

public AccidentArchive(List<Accident> accidents) {
    this.accidents = accidents;
}

So, we are working with streams which I don't understand entirely yet, and I've been stuck in this exercise where I have to make a function that returns a SortedMap<K, V> in which the key has to be the driverAge values and the value has to be a list sorted in descending order of the n lowest vehicleAge values with the same driverAge value:

public SortedMap<Integer, List<Integer>> getNMinVehicleAgesPerDriverAge(Integer n) {
    return getAccidents().stream().
...

I have tried using Collectors.toMap() and Collectors.toList() to somehow make it work, but I have no idea how to do it.

Answer 1

Simplified approach

This problem correlates with the algorithmic question of finding N maximum (or minimum) values via partial sorting. It's implementation using collectors might seem tough, therefore I've decided to introduce a simplified solution.

We can use a flavor of groupingBy() that expects three arguments :

• a classifier function ,
• a supplier mapFactory (which allows to specify the resulting type of map)
• and a downstream collector .

As a downstream collector of groupingBy() we can utilize collectingAndThen with a combination of collectors mapping() and toList() and a function that will sort the whole resulting list mapped to each key , and then will drop unnecessary values retaining only n lowest vehicleAge :

public SortedMap<Integer, List<Integer>> getNMinVehicleAgesPerDriverAge(Integer n) {
    return getAccidents().stream()
        .collect(Collectors.groupingBy(Accident::getDriverAge,
            TreeMap::new,
            Collectors.collectingAndThen(
                Collectors.mapping(Accident::getVehicleAge, Collectors.toList()),
                list -> list.stream()
                    .sorted(Comparator.reverseOrder())
                    .limit(n)
                    .collect(Collectors.toList()))));
}

More performant version

As I've said before, we don't need to sort all values mapped to each key when we need only some of them. When n is dwarfish in comparison to the total size of the list (like 3 to 100,000 for every key) it will cause a serious performance hit.

We can introduce a partial sorting by using PriorityQueue (which is an implementation of the Heap data structure built-in in the JDK).

To enhance the previous solution, we need to replace the downstream collector of groupingBy() you can utilize a combination of mapping() and a custom collector backed by the PriorityQueue which retain only n lowest vehicleAge values associated with each driverAge :

public SortedMap<Integer, List<Integer>> getNMinVehicleAgesPerDriverAge(Integer n) {
    return getAccidents().stream()
        .collect(Collectors.groupingBy(Accident::getDriverAge,
            TreeMap::new,
            Collectors.mapping(Accident::getVehicleAge, 
                getMaxN(n, Comparator.<Integer>reverseOrder()))));
}

The method provided below is responsible for generating a custom collector based on the provided maximum size of the resulting list and a comparator . The logic behind it was explained in great detail in this answer :

public static <T> Collector<T, ?, List<T>> getMaxN(int size, Comparator<T> comparator) {
        
    return Collector.of(
        () -> new PriorityQueue<>(comparator),
        (Queue<T> queue, T next) -> tryAdd(queue, next, comparator, size),
        (Queue<T> left, Queue<T> right) -> {
            right.forEach(next -> tryAdd(left, next, comparator, size));
            return left;
        },
        (Queue<T> queue) -> queue.stream().toList(),
        Collector.Characteristics.UNORDERED);
}
    
public static <T> void tryAdd(Queue<T> queue, T next, Comparator<T> comparator, int size) {
    if (queue.size() == size && comparator.compare(next, queue.element()) < 0) queue.remove(); // if next value is less than the smallest element in the queue and max size has been exceeded the largest element needs to be removed from the queue
    if (queue.size() < size) queue.add(next);
}

By the way, if your assignment doesn't specify the requirement to utilize SortedMap as a return type. It better to use NavigableMap interface, which defines a wider range of methods, instead.