How to get multiple values from an object using a single stream operation?

Question

I want to determine the minimum area required to display a collection of points. The easy way is to loop through the collection like this:

int minX = Integer         


        

Answer 1

Thanks everyone for all the suggestions and answers. This is very helpful and I learned a lot!

I decided to give most of your solutions a go (except for the JDK12 solution). For some of them you already provide me with the code. In addition, I made my own Collector.

class extremesCollector implements Collector, Map> {

    @Override
    public Supplier> supplier() {
        Map map = new HashMap<>();
        map.put("xMin", Integer.MAX_VALUE);
        map.put("yMin", Integer.MAX_VALUE);
        map.put("xMax", Integer.MIN_VALUE);
        map.put("yMax", Integer.MIN_VALUE);
        return () -> map;
    }

    @Override
    public BiConsumer, Point> accumulator() {
        return (a, b) -> {
            a.put("xMin", Math.min(a.get("xMin"), b.x));
            a.put("yMin", Math.min(a.get("yMin"), b.y));
            a.put("xMax", Math.max(a.get("xMax"), b.x));
            a.put("yMax", Math.max(a.get("yMax"), b.y));
        };
    }

    @Override
    public Function, Map> finisher() {
        return Function.identity();
    }

    @Override
    public BinaryOperator> combiner() {
        return (a, b) -> {
            a.put("xMin", Math.min(a.get("xMin"), b.get("xMin")));
            a.put("yMin", Math.min(a.get("yMin"), b.get("yMin")));
            a.put("xMax", Math.max(a.get("xMax"), b.get("xMax")));
            a.put("yMax", Math.max(a.get("yMax"), b.get("yMax")));
            return a;
        };
    }

    @Override
    public Set characteristics() {
        Set characteristics = new HashSet<>();
        characteristics.add(Characteristics.UNORDERED);
        characteristics.add(Characteristics.CONCURRENT);
        characteristics.add(Characteristics.IDENTITY_FINISH);
        return characteristics;
    }
}

Results

I gave all of them a try and compared the results. Good news: for all of them, I got the same result as far as the values are concerned!

With respect to the speed, here is the ranking:

1. for-loops
2. four individual streams
3. stream with self-made Collector
4. parallel stream with self-made Collector
5. statistics approach provided by Andrew Tobilko

Number 2 and 3 are actually very close in terms of speed. The parallel version is probably slower because my dataset is too small.