CODEWITHSUNDEEP
javasorting
callmebob
callmebob

Reputation: 6323

Grouping and Double sorting List<Person>

I went through all the manuals out there and all SO questions but still unable to figure this out...

I have a List (integer represents age):

List<Person> people = Arrays.asList
    (
            new Person("bob", 10),
            new Person("sue", 4),
            new Person("tom", 37),
            new Person("jim", 10),
            new Person("boo", 4),
            new Person("ekk", 53),
            new Person("joe", 10)
    );

I need to:

So using the example above the result would have to be like this: 

{10=[bob, jim, joe],4=[sue, boo], 53=[ekk], 37=[tom]}

What I tried:

I tried with and without streams. I failed on both.

Note: I would lean toward no stream solution, because from my testing of the below code it seems like streams are much slower (I used System.nanotime()). These 3 operations will be done thousands of times each time, so it may make a slight difference.

Using streams here is what I did:

List<List<Person>> grpd = new ArrayList<>
    (
            people.stream()
                    .collect
                            (
                                    groupingBy(Person::getAge, toList())
                            )
                    .values()
    );
grpd = grpd.stream().sorted((a, b) -> Integer.compare(b.size(), a.size())).collect(toList());

No streams approach:

    Map<Integer, List<Person>> grouped = new HashMap<>();
    for (Person person : people)
    {
        if (grouped.containsKey(person._age))
        {
            grouped.get(person._age).add(person);
        } else
        {
            List<Person> p = new ArrayList<>();
            p.add(person);
            grouped.put(person._age, p);
        }
    }


    List<Map.Entry<Integer, List<Person>>> entries = new ArrayList<>(grouped.entrySet());
    Collections.sort(entries, new Comparator<Map.Entry<Integer, List<Person>>>()
    {
        @Override
        public int compare(Map.Entry<Integer, List<Person>> o1, Map.Entry<Integer, List<Person>> o2)
        {
            return Integer.compare(o2.getValue().size(), o1.getValue().size());
        }
    });
    Map<Integer, List<Person>> sortedBySize = new LinkedHashMap<>();
    for (Map.Entry<Integer, List<Person>> entry : entries)
    {
        sortedBySize.put(entry.getKey(), entry.getValue());
    }

Problem: I have no idea how to add the final sort on either case.

public class Person
{
    public String _name;
    public int _age;
    public int getAge() { return _age; }

    public Person(String name, int age)
    {
        _name = name;
        _age = age;
    }

    @Override
    public String toString()
    {
        return _name;
    }
}

Upvotes: 3

Views: 148

Answers (4)

fps
fps

Reputation: 34470

As you've also asked about a non-stream solution, here it is:

Map<Integer, List<Person>> grouped = new HashMap<>();
people.forEach(person -> grouped.computeIfAbsent(
        person.getAge(), 
        k -> new ArrayList<>())
    .add(person));

This groups by age. Now let's sort the entries, first by group size descending, then by age descending:

List<Map.Entry<Integer, List<Person>>> toSort = new ArrayList<>(grouped.entrySet());
toSort.sort(
    Comparator.comparingInt((Map.Entry<Integer, List<Person>> e) -> e.getValue().size())
              .reversed()
              .thenComparingInt(Map.Entry.comparingByKey().reversed()));

Now, toSort is a sorted list of entries. You need to put those entries into a new map:

Map<Integer, List<Person>> sorted = new LinkedHashMap<>();
toSort.forEach(e -> sorted.put(e.getKey(), e.getValue()));

And sorted holds the result you want.

Upvotes: 1

Slaw
Slaw

Reputation: 46181

Since you were also looking for a non-stream solution:

public static Map<Integer, List<Person>> group(List<Person> people) {
  Map<Integer, List<Person>> intermediateGrouping = new HashMap<>();
  for (Person person : people) {
    intermediateGrouping.computeIfAbsent(person.getAge(), k -> new ArrayList<>()).add(person);
  }

  Comparator<Entry<Integer, List<Person>>> byGroupSize = Entry.comparingByValue(Comparator.comparingInt(List::size));
  Comparator<Entry<Integer, List<Person>>> byAge = Entry.comparingByKey();

  List<Entry<Integer, List<Person>>> entries = new ArrayList<>(intermediateGrouping.entrySet());
  entries.sort(byGroupSize.reversed().thenComparing(byAge.reversed()));

  Map<Integer, List<Person>> result = new LinkedHashMap<>(entries.size());
  for (Entry<Integer, List<Person>> entry : entries) {
    result.put(entry.getKey(), entry.getValue());
  }

  return result;
}

Or if you prefer the result to be a List<List<Person>>:

public static List<List<Person>> group(List<Person> people) {
  Map<Integer, List<Person>> intermediateGrouping = new HashMap<>();
  for (Person person : people) {
    intermediateGrouping.computeIfAbsent(person.getAge(), k -> new ArrayList<>()).add(person);
  }

  Comparator<Entry<Integer, List<Person>>> byGroupSize = Entry.comparingByValue(Comparator.comparingInt(List::size));
  Comparator<Entry<Integer, List<Person>>> byAge = Entry.comparingByKey();

  List<Entry<Integer, List<Person>>> entries = new ArrayList<>(intermediateGrouping.entrySet());
  entries.sort(byGroupSize.reversed().thenComparing(byAge.reversed()));

  List<List<Person>> result = new ArrayList<>(entries.size());
  for (Entry<Integer, List<Person>> entry : entries) {
    result.add(entry.getValue());
  }

  return result;
}

Upvotes: 1

Renny
Renny

Reputation: 100

Try modifiying your sort comparator using the below implementation when the sizes are equal for a no streams approach

 Collections.sort(entries, new Comparator<Map.Entry<Integer, List<Person>>>()
        {
            @Override
            public int compare(Map.Entry<Integer, List<Person>> o1, Map.Entry<Integer, List<Person>> o2)
            {
                if(o1.getValue().size()<o2.getValue().size())
                    return 1;
                else if(o1.getValue().size()>o2.getValue().size())
                    return -1;
                else {
                   if( o1.getKey()< o2.getKey())
                            return 1;
                   else if(o1.getKey()>o2.getKey())
                       return -1;
                   else
                       return 0;
                }


            }
        });

Let me know if it works on all your Test Cases

Upvotes: 0

Andy Turner
Andy Turner

Reputation: 140504

Use streams.

First, group them by age:

Map<Integer, List<Person>> groupedByAge =
    people.stream().collect(groupingBy(Person::getAge));

Then sort the entries of this map:

Comparator<Map.Entry<Integer, List<Person>>> byCount = comparingInt(e -> e.getValue().size());
Comparator<Map.Entry<Integer, List<Person>>> byAge = comparingInt(Map.Entry::getKey);
Stream<Map.Entry<Integer, List<Person>>> sorted =
    groupedByAge.entrySet().stream().sorted(byCount.reversed().thenComparing(byAge.reversed()));

Then just get the list out of there:

List<List<Person>> result = sorted.map(Map.Entry::getValue).collect(toList());

(You can put this all into a single expression, but I claim it is more readable broken out like this).

Upvotes: 3

Related Questions