1. Información general
En este artículo, veremos cómo usar HashMap en Java y veremos cómo funciona internamente.
Una clase muy similar a HashMap es Hashtable . Consulte algunos de nuestros otros artículos para obtener más información sobre la clase java.util.Hashtable y las diferencias entre HashMap y Hashtable .
2. Uso básico
Primero veamos qué significa que HashMap es un mapa. Un mapa es un mapeo de clave-valor, lo que significa que cada clave se asigna a exactamente un valor y que podemos usar la clave para recuperar el valor correspondiente de un mapa.
Uno podría preguntarse por qué no simplemente agregar el valor a una lista. ¿Por qué necesitamos un HashMap ? La simple razón es el rendimiento. Si queremos encontrar un elemento específico en una lista, la complejidad del tiempo es O (n) y si la lista está ordenada, será O (log n) usando, por ejemplo, una búsqueda binaria.
La ventaja de un HashMap es que la complejidad de tiempo para insertar y recuperar un valor es O (1) en promedio. Veremos cómo se puede lograr eso más adelante. Primero veamos cómo usar HashMap .
2.1. Preparar
Creemos una clase simple que usaremos a lo largo del artículo:
public class Product { private String name; private String description; private List tags; // standard getters/setters/constructors public Product addTagsOfOtherProdcut(Product product) { this.tags.addAll(product.getTags()); return this; } }2.2. Poner
Ahora podemos crear un HashMap con la clave de tipo String y elementos de tipo Product :
Map productsByName = new HashMap(); Y agregue productos a nuestro HashMap :
Product eBike = new Product("E-Bike", "A bike with a battery"); Product roadBike = new Product("Road bike", "A bike for competition"); productsByName.put(eBike.getName(), eBike); productsByName.put(roadBike.getName(), roadBike); 2.3. Obtener
Podemos recuperar un valor del mapa por su clave:
Product nextPurchase = productsByName.get("E-Bike"); assertEquals("A bike with a battery", nextPurchase.getDescription());Si intentamos encontrar un valor para una clave que no existe en el mapa, obtendremos un valor nulo :
Product nextPurchase = productsByName.get("Car"); assertNull(nextPurchase);Y si insertamos un segundo valor con la misma clave, solo obtendremos el último valor insertado para esa clave:
Product newEBike = new Product("E-Bike", "A bike with a better battery"); productsByName.put(newEBike.getName(), newEBike); assertEquals("A bike with a better battery", productsByName.get("E-Bike").getDescription());2.4. Nulo como clave
HashMap también nos permite tener null como clave:
Product defaultProduct = new Product("Chocolate", "At least buy chocolate"); productsByName.put(null, defaultProduct); Product nextPurchase = productsByName.get(null); assertEquals("At least buy chocolate", nextPurchase.getDescription());2.5. Valores con la misma clave
Además, podemos insertar el mismo objeto dos veces con una clave diferente:
productsByName.put(defaultProduct.getName(), defaultProduct); assertSame(productsByName.get(null), productsByName.get("Chocolate"));2.6. Eliminar un valor
Podemos eliminar una asignación de clave-valor del HashMap :
productsByName.remove("E-Bike"); assertNull(productsByName.get("E-Bike"));2.7. Compruebe si existe una clave o valor en el mapa
Para verificar si una clave está presente en el mapa, podemos usar el método containsKey () :
productsByName.containsKey("E-Bike");O, para verificar si un valor está presente en el mapa, podemos usar el método containsValue () :
productsByName.containsValue(eBike);Ambas llamadas a métodos devolverán verdadero en nuestro ejemplo. Aunque son muy similares, existe una diferencia importante en el rendimiento entre estas dos llamadas a métodos. La complejidad para verificar si existe una clave es O (1) , mientras que la complejidad para verificar un elemento es O (n), ya que es necesario recorrer todos los elementos del mapa.
2.8. Iterando sobre un HashMap
Hay tres formas básicas de iterar sobre todos los pares clave-valor en un HashMap .
Podemos iterar sobre el conjunto de todas las claves:
for(String key : productsByName.keySet()) { Product product = productsByName.get(key); }O podemos iterar sobre el conjunto de todas las entradas:
for(Map.Entry entry : productsByName.entrySet()) { Product product = entry.getValue(); String key = entry.getKey(); //do something with the key and value }Finalmente, podemos iterar sobre todos los valores:
List products = new ArrayList(productsByName.values());3. La clave
Podemos usar cualquier clase como clave en nuestro HashMap . Sin embargo, para que el mapa funcione correctamente, necesitamos proporcionar una implementación para equals () y hashCode (). Digamos que queremos tener un mapa con el producto como clave y el precio como valor:
HashMap priceByProduct = new HashMap(); priceByProduct.put(eBike, 900);Implementemos los métodos equals () y hashCode () :
@Override public boolean equals(Object o) { if (this == o) { return true; } if (o == null || getClass() != o.getClass()) { return false; } Product product = (Product) o; return Objects.equals(name, product.name) && Objects.equals(description, product.description); } @Override public int hashCode() { return Objects.hash(name, description); }Tenga en cuenta que hashCode () y equals () deben anularse solo para las clases que queremos usar como claves de mapa, no para las clases que solo se usan como valores en un mapa. Veremos por qué esto es necesario en la sección 5 de este artículo.
4. Additional Methods as of Java 8
Java 8 added several functional-style methods to HashMap. In this section, we'll look at some of these methods.
For each method, we'll look at two examples. The first example shows how to use the new method, and the second example shows how to achieve the same in earlier versions of Java.
As these methods are quite straightforward, we won't look at more detailed examples.
4.1. forEach()
The forEach method is the functional-style way to iterate over all elements in the map:
productsByName.forEach( (key, product) -> { System.out.println("Key: " + key + " Product:" + product.getDescription()); //do something with the key and value }); Prior to Java 8:
for(Map.Entry entry : productsByName.entrySet()) { Product product = entry.getValue(); String key = entry.getKey(); //do something with the key and value }Our article Guide to the Java 8 forEach covers the forEach loop in greater detail.
4.2. getOrDefault()
Using the getOrDefault() method, we can get a value from the map or return a default element in case there is no mapping for the given key:
Product chocolate = new Product("chocolate", "something sweet"); Product defaultProduct = productsByName.getOrDefault("horse carriage", chocolate); Product bike = productsByName.getOrDefault("E-Bike", chocolate);Prior to Java 8:
Product bike2 = productsByName.containsKey("E-Bike") ? productsByName.get("E-Bike") : chocolate; Product defaultProduct2 = productsByName.containsKey("horse carriage") ? productsByName.get("horse carriage") : chocolate; 4.3. putIfAbsent()
With this method, we can add a new mapping, but only if there is not yet a mapping for the given key:
productsByName.putIfAbsent("E-Bike", chocolate); Prior to Java 8:
if(productsByName.containsKey("E-Bike")) { productsByName.put("E-Bike", chocolate); }Our article Merging Two Maps with Java 8 takes a closer look at this method.
4.4. merge()
And with merge(), we can modify the value for a given key if a mapping exists, or add a new value otherwise:
Product eBike2 = new Product("E-Bike", "A bike with a battery"); eBike2.getTags().add("sport"); productsByName.merge("E-Bike", eBike2, Product::addTagsOfOtherProdcut);Prior to Java 8:
if(productsByName.containsKey("E-Bike")) { productsByName.get("E-Bike").addTagsOfOtherProdcut(eBike2); } else { productsByName.put("E-Bike", eBike2); } 4.5. compute()
With the compute() method, we can compute the value for a given key:
productsByName.compute("E-Bike", (k,v) -> { if(v != null) { return v.addTagsOfOtherProdcut(eBike2); } else { return eBike2; } });Prior to Java 8:
if(productsByName.containsKey("E-Bike")) { productsByName.get("E-Bike").addTagsOfOtherProdcut(eBike2); } else { productsByName.put("E-Bike", eBike2); } It's worth noting that the methods merge() and compute() are quite similar. The compute() method accepts two arguments: the key and a BiFunction for the remapping. And merge() accepts three parameters: the key, a default value to add to the map if the key doesn't exist yet, and a BiFunction for the remapping.
5. HashMap Internals
In this section, we'll look at how HashMap works internally and what are the benefits of using HashMap instead of a simple list, for example.
As we've seen, we can retrieve an element from a HashMap by its key. One approach would be to use a list, iterate over all elements, and return when we find an element for which the key matches. Both the time and space complexity of this approach would be O(n).
With HashMap, we can achieve an average time complexity of O(1) for the put and get operations and space complexity of O(n). Let's see how that works.
5.1. The Hash Code and Equals
Instead of iterating over all its elements, HashMap attempts to calculate the position of a value based on its key.
The naive approach would be to have a list that can contain as many elements as there are keys possible. As an example, let's say our key is a lower-case character. Then it's sufficient to have a list of size 26, and if we want to access the element with key ‘c', we'd know that it's the one at position 3, and we can retrieve it directly.
However, this approach would not be very effective if we have a much bigger keyspace. For example, let's say our key was an integer. In this case, the size of the list would have to be 2,147,483,647. In most cases, we would also have far fewer elements, so a big part of the allocated memory would remain unused.
HashMap stores elements in so-called buckets and the number of buckets is called capacity.
When we put a value in the map, the key's hashCode() method is used to determine the bucket in which the value will be stored.
To retrieve the value, HashMap calculates the bucket in the same way – using hashCode(). Then it iterates through the objects found in that bucket and use key's equals() method to find the exact match.
5.2. Keys' Immutability
In most cases, we should use immutable keys. Or at least, we must be aware of the consequences of using mutable keys.
Let's see what happens when our key changes after we used it to store a value in a map.
For this example, we'll create the MutableKey:
public class MutableKey { private String name; // standard constructor, getter and setter @Override public boolean equals(Object o) { if (this == o) { return true; } if (o == null || getClass() != o.getClass()) { return false; } MutableKey that = (MutableKey) o; return Objects.equals(name, that.name); } @Override public int hashCode() { return Objects.hash(name); } }And here goes the test:
MutableKey key = new MutableKey("initial"); Map items = new HashMap(); items.put(key, "success"); key.setName("changed"); assertNull(items.get(key));As we can see, we're no longer able to get the corresponding value once the key has changed, instead, null is returned. This is because HashMap is searching in the wrong bucket.
The above test case may be surprising if we don't have a good understanding of how HashMap works internally.
5.3. Collisions
For this to work correctly, equal keys must have the same hash, however, different keys can have the same hash. If two different keys have the same hash, the two values belonging to them will be stored in the same bucket. Inside a bucket, values are stored in a list and retrieved by looping over all elements. The cost of this is O(n).
As of Java 8 (see JEP 180), the data structure in which the values inside one bucket are stored is changed from a list to a balanced tree if a bucket contains 8 or more values, and it's changed back to a list if, at some point, only 6 values are left in the bucket. This improves the performance to be O(log n).
5.4. Capacity and Load Factor
To avoid having many buckets with multiple values, the capacity is doubled if 75% (the load factor) of the buckets become non-empty. The default value for the load factor is 75%, and the default initial capacity is 16. Both can be set in the constructor.
5.5. Summary of put and get Operations
Let's summarize how the put and get operations work.
When we add an element to the map,HashMap calculates the bucket. If the bucket already contains a value, the value is added to the list (or tree) belonging to that bucket. If the load factor becomes bigger than the maximum load factor of the map, the capacity is doubled.
When we want to get a value from the map,HashMap calculates the bucket and gets the value with the same key from the list (or tree).
6. Conclusion
En este artículo, vimos cómo usar un HashMap y cómo funciona internamente. Junto con ArrayList , HashMap es una de las estructuras de datos más utilizadas en Java, por lo que es muy útil tener un buen conocimiento de cómo usarlo y cómo funciona bajo el capó. Nuestro artículo The Java HashMap Under the Hood cubre los aspectos internos de HashMap con más detalle.
Como de costumbre, el código fuente completo está disponible en Github.