The primary advantage of lambdas over anonymous classes is that they are more succinct. Java provides a way to generate function objects even more succinct than lambdas: method references. Here is a code snippet from a program that maintains a map from arbitrary keys to Integer values. If the value is interpreted as a count of the number of instances of the key, then the program is a multiset implementation. The function of the code snippet is to associate the number 1 with the key if it is not in the map and to increment the associated value if the key is already present:
map.merge(key, 1, (count, incr) -> count + incr);
Note that this code uses the merge method, which was added to the Map interface in Java 8. If no mapping is present for the given key, the method simply inserts the given value; if a mapping is already present, merge applies the given function to the current value and the given value and overwrites the current value with the result. This code represents a typical use case for the merge method.
The code reads nicely, but there’s still some boilerplate. The parameters count and incr don’t add much value, and they take up a fair amount of space. Really, all the lambda tells you is that the function returns the sum of its two arguments. As of Java 8, Integer (and all the other boxed numerical primitive types) provides a static method sum that does exactly the same thing. We can simply pass a reference to this method and get the same result with less visual clutter:
map.merge(key, 1, Integer::sum);
The more parameters a method has, the more boilerplate you can eliminate with a method reference. In some lambdas, however, the parameter names you choose provide useful documentation, making the lambda more readable and maintainable than a method reference, even if the lambda is longer.
There’s nothing you can do with a method reference that you can’t also do with a lambda (with one obscure exception—see JLS, 9.9-2 if you’re curious). That said, method references usually result in shorter, clearer code. They also give you an out if a lambda gets too long or complex: You can extract the code from the lambda into a new method and replace the lambda with a reference to that method. You can give the method a good name and document it to your heart’s content.
If you’re programming with an IDE, it will offer to replace a lambda with a method reference wherever it can. You should usually, but not always, take the IDE up on the offer. Occasionally, a lambda will be more succinct than a method reference. This happens most often when the method is in the same class as the lambda. For example, consider this snippet, which is presumed to occur in a class named GoshThisClassNameIsHumongous:
service.execute(GoshThisClassNameIsHumongous::action);
The lambda equivalent looks like this:
service.execute(() -> action());
The snippet using the method reference is neither shorter nor clearer than the snippet using the lambda, so prefer the latter. Along similar lines, the Function interface provides a generic static factory method to return the identity function, Function.identity(). It’s typically shorter and cleaner not to use this method but to code the equivalent lambda inline: x -> x.
Many method references refer to static methods, but there are four kinds that do not. Two of them are bound and unbound instance method references. In bound references, the receiving object is specified in the method reference. Bound references are similar in nature to static references: the function object takes the same arguments as the referenced method. In unbound references, the receiving object is specified when the function object is applied, via an additional parameter before the method’s declared parameters. Unbound references are often used as mapping and filter functions in stream pipelines (Item 45). Finally, there are two kinds of constructor references, for classes and arrays. Constructor references serve as factory objects. All five kinds of method references are summarized in the table below:
| Method Ref Type | Example | Lambda Equivalent |
|---|---|---|
| Static | Integer::parseInt | str -> |
| Bound | Instant.now()::isAfter | Instant then =Instant.now(); t ->then.isAfter(t) |
| Unbound | String::toLowerCase | str ->str.toLowerCase() |
| Class Constructor | TreeMap<K,V>::new | () -> new TreeMap<K,V> |
| Array Constructor | int[]::new | len -> new int[len] |
In summary, method references often provide a more succinct alternative to lambdas. Where method references are shorter and clearer, use them; where they aren’t, stick with lambdas.