CODEWITHSUNDEEP
javafunctional-programmingjava-8
jakew
jakew

Reputation: 59

Passing function code using java 8 supplier

I am trying to understand lambdas in Java 8. I have this very simple example using a functional interface.

This is not a working example and I understand that I can directly call method1 and method2 rather than using getValue() and Supplier, but, I wanted to know if there are any issues with calling method1 and method2 using supplier.get().

public <T> T mytest(SomeObject var) {
    //extract values from someobject
    if(somecondition) {
        getValue(() ->  method2(obj5instance, obj6instance));
    } else {
        getValue(() ->  method1("somestr", obj1instance, obj2instance, obj3instance));
    }
}

private String method1(String s1, Obj1 s2, Obj2 s3, Obj3 s4){
    //network call here to compute return value
    return "computedmethod1";
}

private String method2(Obj5 s2, Obj6 s3){
    //network call here to compute return value
    return "computedmethod2";
}

public <T> T getValue(Supplier<T> supplier){    
    //this will call method1 or method2
    //wrap the return value from supplier.get and return
    return supplier.get();
}

Update based on Andrew's comment.

That makes a lot of sense, thanks. I am having difficult wrapping my head around the concept of supplier.

C1.java 

public C3 m1(Obj obj){      
  Supplier<String> mysupplier = () -> foo(arg1, arg2, arg3);       //foo is an expensive call
return someFn(mysupplier);    
}

How does supplier work internally? When the supplier is created, does it keep a reference to foo() and the arguments passed to foo? 

public C3 someFn(Supplier<String> supplier){   
//this returns an instance of C3 that has a method lazyCall() which runs supplier.get()   
}

C2.java 

C3 c3 = c1.m1(obj);

If and when c3.lazyCall() is called, it runs the supplier.get() and foo is run. Is this a valid use of supplier?

Upvotes: 2

Views: 3564

Answers (1)

Andrew
Andrew

Reputation: 49606

One of the advantages of the Supplier interface is providing a value in a lazy way.

We try to delay a supplier.get() call as far as possible, because it's better to execute a resource-consuming operation (if you have such) at the moment when it really needs to be performed avoiding cases where the result of the operation can be ignored or might be skipped.

Here, the wrapper over a Supplier is senseless, because there are no conditions where it may not get called. Also, I don't see any reasons why these methods should be executed lazily. Though, I don't see a whole processing too.

EDIT:

How does supplier work internally? When the supplier is created, does it keep a reference to foo() and the arguments passed to foo?

Think about a lambda expression like a shortcut to anonymous class declaration (as it actually is) with overriding a single method. In the lambda body, you simply write method code which is gonna be executed when it gets invoked.

No, it doesn't keep references, because it has a direct access to them (with some limitations like effectively final variables).

Upvotes: 4

Related Questions