Sunday, 26 February 2017

Working with primitives in Streams

Background

In one of the previous posts on Understanding Java 8 Stream API we saw basics of working with primitives. To reiterate - there are 3 types of primitive streams -
  1. IntStream: Used for the primitive types int, short, byte, and char
  2. LongStream: Used for the primitive type long
  3. DoubleStream: Used for the primitive types double and float
These primitives are special they got their own Optional classes , their own functional interfaces. With our previous understanding in place we will see this new working with primitives.


Why primitive Streams?

Lets say we have a list of integers and we want to compute its average. How would you do it with streams.

    public static void main(String[] args) {
        List<Integer> myIntList = Arrays.asList(1,2,3,4,5);
        Optional<Integer> sum = myIntList.stream().reduce((x,y) -> x + y);
        long count  = myIntList.stream().count();
        double avg = sum.get()/count;
        System.out.println(avg);
    }

Output : 3.0

Well sure based on what we know so far we aced this. But do you see any drawback in above approach?
  • You are creating stream twice from the list - once to get sum and once to get count.
  • Then you are dividing it to get the average.
Though this works Java provides more optimized solution and this is where IntStream comes into picture.

    public static void main(String[] args) {
        List<Integer> myIntList = Arrays.asList(1,2,3,4,5);
        OptionalDouble intAvg = myIntList.stream().mapToInt(x->x).average();
        System.out.println(intAvg.getAsDouble());
    }


Output : 3.0

That's the one liner with IntStream. mapToInt() intermediate operation gives you a IntStream which has predefined operations like sum(), average() etc. Couple of points to note -
  • mapToInt() intermediate method returned IntStream
  •  average() terminal operation on IntStream returned a OptionalDouble. This is a new class. Notice it is not Optional<Double> but entirely a new class. 
  • Also note the method to get double value from  OptionalDouble class. It is getAsDouble() and not get() like you would have seen in Optional class.
  • Recollect out discussion before primitives have their own streams eg. IntStream , own Optional classes eg. OptionalInt and even functional interfaces eg. BooleanPredicate. But that's for later.
  • sum() on IntStream returns int instead of OptionalInt because if there are no elements in the Stream it would return 0.
    public static void main(String[] args) {
        List<Integer> myIntList = Arrays.asList(1,2,3,4,5);
        int sum = myIntList.stream().mapToInt(x->x).sum();
        System.out.println(sum);
    } 


Output : 15

NOTE : If you want Object stream back from primitive streams you can use mapToObj intermediate method.

Creating primitive streams

Now that we have basics out of our way let see how can we create a primitive stream. One method we already saw by using mapToInt() method on a normal Stream. Similarly you have mapToLong() and mapToDouble() methods to get LongStream and DoubleStream respectively. 

Other normal methods that we say in Stream class like of() or generate() or iterate() still apply to primitive streams as well.

    public static void main(String[] args) {
        System.out.println("of");
        IntStream.of(1,2,3,4,5).forEach(System.out::print);
        System.out.println("\ngenerate");
        IntStream.generate(() -> 1).limit(4).forEach(System.out::print);
        System.out.println("\niterate");
        IntStream.iterate(1,(x) -> x+1).limit(4).forEach(System.out::print);
    }

Output :
of
12345
generate
1111
iterate
1234

Some other methods are 
  • range(int startInclusive, int endExclusive)
  • rangeClosed(int startInclusive, int endInclusive)
    public static void main(String[] args) {
        System.out.println("range");
        IntStream.range(1, 5).forEach(System.out::print);
        System.out.println("\nrangeClosed");
        IntStream.rangeClosed(1, 5).forEach(System.out::print);
    } 


Output :
range
1234
rangeClosed
12345

Example demonstrates for InStream but same applies for LongStream and DoubleStream as well.

NOTE : Notice the range limits. For range end limit is exclusive where as for rangeClosed it is inclusive.


Understanding summaryStatistics

Now lets say we have an IntStream and we need to to output difference between the max element and min element in the stream. How would you do that?

    public static void main(String[] args) {
        List<Integer> myList = Arrays.asList(1,2,3,4,5);
        int max = myList.stream().mapToInt(x -> x).max().getAsInt();
        int min = myList.stream().mapToInt(x -> x).min().getAsInt();
        int difference = max - min;
        System.out.println("Difference : " + difference);
    }

Output :
Difference : 4
which is expected since 5-1 = 4.
But again we need to create stream twice and need two terminal operations to get result. Java helps us here too with summaryStatistics method.

    public static void main(String[] args) {
        List<Integer> myList = Arrays.asList(1,2,3,4,5);
        IntSummaryStatistics intSummaryStatistics = myList.stream().mapToInt(x -> x).summaryStatistics();
        int difference = intSummaryStatistics.getMax() - intSummaryStatistics.getMin();
        System.out.println("Difference : " + difference);
    }

Output :
Difference : 4

Related Links

No comments:

Post a comment

t> UA-39527780-1 back to top