Tri en Java

1. Vue d'ensemble

Cet article va illustrer comment appliquer le tri à Array , List , Set et Map dans Java 7 et Java 8.

2. Tri avec matrice

Commençons par trier les tableaux d'entiers en utilisant la méthode Arrays.sort () .

Nous définirons les tableaux int suivants dans une méthode @Before jUnit:

@Before public void initVariables () { toSort = new int[] { 5, 1, 89, 255, 7, 88, 200, 123, 66 }; sortedInts = new int[] {1, 5, 7, 66, 88, 89, 123, 200, 255}; sortedRangeInts = new int[] {5, 1, 89, 7, 88, 200, 255, 123, 66}; ... }

2.1. Tri d'un tableau complet

Utilisons maintenant la simple API Array.sort () :

@Test public void givenIntArray_whenUsingSort_thenSortedArray() { Arrays.sort(toSort); assertTrue(Arrays.equals(toSort, sortedInts)); }

Le tableau non trié est maintenant entièrement trié:

[1, 5, 7, 66, 88, 89, 123, 200, 255]

Comme mentionné dans le JavaDoc officiel, Arrays.sort utilise Quicksort à double pivot sur les primitives . Il offre des performances O (n log (n)) et est généralement plus rapide que les implémentations Quicksort traditionnelles (à un pivot). Cependant, il utilise une implémentation itérative stable et adaptative de l'algorithme de tri de fusion pour Array of Objects.

2.2. Tri d'une partie d'un tableau

Arrays.sort a une autre API de tri - dont nous parlerons ici:

Arrays.sort(int[] a, int fromIndex, int toIndex)

Cela ne triera qu'une partie du tableau, entre les deux index.

Jetons un coup d'œil à un exemple rapide:

@Test public void givenIntArray_whenUsingRangeSort_thenRangeSortedArray() { Arrays.sort(toSort, 3, 7); assertTrue(Arrays.equals(toSort, sortedRangeInts)); }

Le tri sera effectué uniquement sur les éléments de sous-tableau suivants ( toIndex serait exclusif):

[255, 7, 88, 200]

Le sous-tableau trié résultant inclus avec le tableau principal serait:

[5, 1, 89, 7, 88, 200, 255, 123, 66]

2.3. Java 8 Arrays.sort et Arrays.parallelSort

Java 8 est livré avec une nouvelle API - parallelSort - avec une signature similaire à l' API Arrays.sort () :

@Test public void givenIntArray_whenUsingParallelSort_thenArraySorted() { Arrays.parallelSort(toSort); assertTrue(Arrays.equals(toSort, sortedInts)); }

Dans les coulisses de parallelSort (), il divise le tableau en différents sous-tableaux (selon la granularité de l'algorithme de parallelSort ). Chaque sous-tableau est trié avec Arrays.sort () dans différents threads afin que le tri puisse être exécuté de manière parallèle et est finalement fusionné sous forme de tableau trié.

Notez que le pool commun ForJoin est utilisé pour exécuter ces tâches parallèles, puis pour fusionner les résultats.

Le résultat de Arrays.parallelSort sera le même que Array.sort bien sûr, il s'agit simplement de tirer parti du multi-threading.

Enfin, il existe également des variantes similaires d'API Arrays.sort dans Arrays.parallelSort :

Arrays.parallelSort (int [] a, int fromIndex, int toIndex);

3. Tri d'une liste

Utilisons maintenant l' API Collections.sort () dans java.utils.Collections - pour trier une liste d'entiers:

@Test public void givenList_whenUsingSort_thenSortedList() { List toSortList = Ints.asList(toSort); Collections.sort(toSortList); assertTrue(Arrays.equals(toSortList.toArray(), ArrayUtils.toObject(sortedInts))); }

La liste avant le tri contiendra les éléments suivants:

[5, 1, 89, 255, 7, 88, 200, 123, 66]

Et bien entendu, après le tri:

[1, 5, 7, 66, 88, 89, 123, 200, 255]

Comme mentionné dans Oracle JavaDoc for Collections.Sort , il utilise un tri de fusion modifié et offre des performances garanties n log (n) .

4. Tri d'un ensemble

Ensuite, utilisons Collections.sort () pour trier un LinkedHashSet .

Nous utilisons le LinkedHashSet car il maintient l'ordre d'insertion.

Remarquez comment, pour utiliser l' API de tri dans les collections , nous emballons d'abord l'ensemble dans une liste :

@Test public void givenSet_whenUsingSort_thenSortedSet() { Set integersSet = new LinkedHashSet(Ints.asList(toSort)); Set descSortedIntegersSet = new LinkedHashSet( Arrays.asList(new Integer[] {255, 200, 123, 89, 88, 66, 7, 5, 1})); List list = new ArrayList(integersSet); Collections.sort(Comparator.reverseOrder()); integersSet = new LinkedHashSet(list); assertTrue(Arrays.equals( integersSet.toArray(), descSortedIntegersSet.toArray())); }

La méthode Comparator.reverseOrder () inverse l'ordre imposé par l'ordre naturel.

5. Carte de tri

Dans cette section, nous allons commencer à examiner le tri d'une carte - à la fois par clés et par valeurs.

Définissons d'abord la carte que nous allons trier:

@Before public void initVariables () { .... HashMap map = new HashMap(); map.put(55, "John"); map.put(22, "Apple"); map.put(66, "Earl"); map.put(77, "Pearl"); map.put(12, "George"); map.put(6, "Rocky"); .... }

5.1. Tri de la carte par clés

We'll now extract keys and values entries from the HashMap and sort it based on the values of the keys in this example:

@Test public void givenMap_whenSortingByKeys_thenSortedMap() { Integer[] sortedKeys = new Integer[] { 6, 12, 22, 55, 66, 77 }; List
    
      entries = new ArrayList(map.entrySet()); Collections.sort(entries, new Comparator
     
      () { @Override public int compare( Entry o1, Entry o2) { return o1.getKey().compareTo(o2.getKey()); } }); Map sortedMap = new LinkedHashMap(); for (Map.Entry entry : entries) { sortedMap.put(entry.getKey(), entry.getValue()); } assertTrue(Arrays.equals(sortedMap.keySet().toArray(), sortedKeys)); }
     
    

Note how we used the LinkedHashMap while copying the sorted Entries based on keys (because HashSet doesn't guarantee the order of keys).

The Map before sorting :

[Key: 66 , Value: Earl] [Key: 22 , Value: Apple] [Key: 6 , Value: Rocky] [Key: 55 , Value: John] [Key: 12 , Value: George] [Key: 77 , Value: Pearl]

The Map after sorting by keys:

[Key: 6 , Value: Rocky] [Key: 12 , Value: George] [Key: 22 , Value: Apple] [Key: 55 , Value: John] [Key: 66 , Value: Earl] [Key: 77 , Value: Pearl] 

5.2. Sorting Map by Values

Here we will be comparing values of HashMap entries for sorting based on values of HashMap:

@Test public void givenMap_whenSortingByValues_thenSortedMap() { String[] sortedValues = new String[] { "Apple", "Earl", "George", "John", "Pearl", "Rocky" }; List
    
      entries = new ArrayList(map.entrySet()); Collections.sort(entries, new Comparator
     
      () { @Override public int compare( Entry o1, Entry o2) { return o1.getValue().compareTo(o2.getValue()); } }); Map sortedMap = new LinkedHashMap(); for (Map.Entry entry : entries) { sortedMap.put(entry.getKey(), entry.getValue()); } assertTrue(Arrays.equals(sortedMap.values().toArray(), sortedValues)); }
     
    

The Map before sorting:

[Key: 66 , Value: Earl] [Key: 22 , Value: Apple] [Key: 6 , Value: Rocky] [Key: 55 , Value: John] [Key: 12 , Value: George] [Key: 77 , Value: Pearl]

The Map after sorting by values:

[Key: 22 , Value: Apple] [Key: 66 , Value: Earl] [Key: 12 , Value: George] [Key: 55 , Value: John] [Key: 77 , Value: Pearl] [Key: 6 , Value: Rocky]

6. Sorting Custom Objects

Let's now work with a custom object:

public class Employee implements Comparable { private String name; private int age; private double salary; public Employee(String name, int age, double salary) { ... } // standard getters, setters and toString }

We'll be using the following Employee Array for sorting example in the following sections:

@Before public void initVariables () { .... employees = new Employee[] { new Employee("John", 23, 5000), new Employee("Steve", 26, 6000), new Employee("Frank", 33, 7000), new Employee("Earl", 43, 10000), new Employee("Jessica", 23, 4000), new Employee("Pearl", 33, 6000)}; employeesSorted = new Employee[] { new Employee("Earl", 43, 10000), new Employee("Frank", 33, 70000), new Employee("Jessica", 23, 4000), new Employee("John", 23, 5000), new Employee("Pearl", 33, 4000), new Employee("Steve", 26, 6000)}; employeesSortedByAge = new Employee[] { new Employee("John", 23, 5000), new Employee("Jessica", 23, 4000), new Employee("Steve", 26, 6000), new Employee("Frank", 33, 70000), new Employee("Pearl", 33, 4000), new Employee("Earl", 43, 10000)}; }

We can sort arrays or collections of custom objects either:

  1. in the natural order (Using the Comparable Interface) or
  2. in the order provided by a ComparatorInterface

6.1. Using Comparable

The natural order in java means an order in which primitive or Object should be orderly sorted in a given array or collection.

Both java.util.Arrays and java.util.Collections have a sort() method, and It's highly recommended that natural orders should be consistent with the semantics of equals.

In this example, we will consider employees with the same name as equal:

@Test public void givenEmpArray_SortEmpArray_thenSortedArrayinNaturalOrder() { Arrays.sort(employees); assertTrue(Arrays.equals(employees, employeesSorted)); }

You can define the natural order for elements by implementing a Comparable interface which has compareTo() method for comparing current object and object passed as an argument.

To understand this clearly, let's see an example Employee class which implements Comparable Interface:

public class Employee implements Comparable { ... @Override public boolean equals(Object obj) { return ((Employee) obj).getName().equals(getName()); } @Override public int compareTo(Object o) { Employee e = (Employee) o; return getName().compareTo(e.getName()); } }

Generally, the logic for comparison will be written the method compareTo. Here we are comparing the employee order or name of the employee field. Two employees will be equal if they have the same name.

Now when Arrays.sort(employees); is called in the above code, we now know what is the logic and order which goes in sorting the employees as per the age :

[("Earl", 43, 10000),("Frank", 33, 70000), ("Jessica", 23, 4000), ("John", 23, 5000),("Pearl", 33, 4000), ("Steve", 26, 6000)]

We can see the array is sorted by name of the employee – which now becomes a natural order for Employee Class.

6.2. Using Comparator

Now, let's sort the elements using a Comparator interface implementation – where we pass the anonymous inner class on-the-fly to the Arrays.sort() API:

@Test public void givenIntegerArray_whenUsingSort_thenSortedArray() { Integer [] integers = ArrayUtils.toObject(toSort); Arrays.sort(integers, new Comparator() { @Override public int compare(Integer a, Integer b) { return Integer.compare(a, b); } }); assertTrue(Arrays.equals(integers, ArrayUtils.toObject(sortedInts))); }

Now lets sort employees based on salary – and pass in another comparator implementation:

Arrays.sort(employees, new Comparator() { @Override public int compare(Employee o1, Employee o2) { return Double.compare(o1.getSalary(), o2.getSalary()); } });

The sorted Employees arrays based on salary will be:

[(Jessica,23,4000.0), (John,23,5000.0), (Pearl,33,6000.0), (Steve,26,6000.0), (Frank,33,7000.0), (Earl,43,10000.0)] 

Note that we can use Collections.sort() in a similar fashion to sort List and Set of Objects in Natural or Custom order as described above for Arrays.

7. Sorting With Lambdas

Start with Java 8, we can use Lambdas to implement the Comparator Functional Interface.

You can have a look at the Lambdas in Java 8 writeup to brush up on the syntax.

Let's replace the old comparator:

Comparator c = new Comparator() { @Override public int compare(Integer a, Integer b) { return Integer.compare(a, b); } }

With the equivalent implementation, using Lambda expression:

Comparator c = (a, b) -> Integer.compare(a, b);

Finally, let's write the test:

@Test public void givenArray_whenUsingSortWithLambdas_thenSortedArray() { Integer [] integersToSort = ArrayUtils.toObject(toSort); Arrays.sort(integersToSort, (a, b) -> { return Integer.compare(a, b); }); assertTrue(Arrays.equals(integersToSort, ArrayUtils.toObject(sortedInts))); }

As you can see, a much cleaner and more concise logic here.

8. Using Comparator.comparing and Comparator.thenComparing

Java 8 comes with two new APIs useful for sorting – comparing() and thenComparing() in the Comparator interface.

These are quite handy for the chaining of multiple conditions of the Comparator.

Let's consider a scenario where we may want to compare Employee by age and then by name:

@Test public void givenArrayObjects_whenUsingComparing_thenSortedArrayObjects() { List employeesList = Arrays.asList(employees); employees.sort(Comparator.comparing(Employee::getAge)); assertTrue(Arrays.toString(employees.toArray()) .equals(sortedArrayString)); }

In this example, Employee::getAge is the sorting key for Comparator interface implementing a functional interface with compare function.

Here's the array of Employees after sorting:

[(John,23,5000.0), (Jessica,23,4000.0), (Steve,26,6000.0), (Frank,33,7000.0), (Pearl,33,6000.0), (Earl,43,10000.0)]

Here the employees are sorted based on age.

We can see John and Jessica are of same age – which means that the order logic should now take their names into account- which we can achieve with thenComparing():

... employees.sort(Comparator.comparing(Employee::getAge) .thenComparing(Employee::getName)); ... 

After sorting with above code snippet, the elements in employee array would be sorted as:

[(Jessica,23,4000.0), (John,23,5000.0), (Steve,26,6000.0), (Frank,33,7000.0), (Pearl,33,6000.0), (Earl,43,10000.0) ]

Thus comparing() and thenComparing() definitely make more complex sorting scenarios a lot cleaner to implement.

9. Conclusion

In this article, we saw how we can apply sorting to Array, List, Set, and Map.

Nous avons également vu une brève introduction sur la façon dont les fonctionnalités de Java 8 pourraient être utiles dans le tri, comme l'utilisation de Lambdas, Compare () , puisComparing () et parallelSort () .

Tous les exemples utilisés dans l'article sont disponibles à l'adresse over sur GitHub.