Tohuwabohu excorcism

New features:

• Methods as first class citizens
• Lambdas
  • Passing code to methods (behavior parametrization)
• Default methods in Interfaces
• Streams
• Nashorn
• New Date API
• Optional

Which type do new Java8 expressions have?

		
		// Function. Gets Ranking. Returns Integer.
		Function<Ranking, Integer> f = ranking -> ranking.value;
		
		// Consumer - accepts sinle parameter. Returns nothing
		Consumer<Ranking> c = Ranking:: echo; 
		
		// Predicate - boolean returning function
		Predicate<Ranking> isLarger2 = ranking -> ranking.value > 2;




	// MODELS
	
	static class Ranking{
		int value;

		public static void echo(Object value){
			System.out.println(String.format("Value is '%s'",value));
		}
	}

You can pass a method as a function-parameter. For that the method-signature must define a functional interface as a paramter.

• functional Interfaces are located in package java.util.function
  • e.g. Consumer<A>
  • Biconsumer<A,B> …

Object o = () -> {return 2;} // invalid Object in not a functional interface

Functional Interfaces may be annotated with @FunctionalInterface to clarify what they are!

The generic functions have following names:

class Test{
public static void main(String[] args){

// passing myMethod, which may be used inside fun
Test.function(Test::myMethod);

}


// As defined by class FunctionalInterface - you can pass any function in here, 
// which takes String as parameter AND
// which returns boolean
public void function(FunctionalInterface interface){
}

static interface FunctionalInterface{
 boolean function(String parameter); // defines a single function operating on String, returning boolean
}

static class MyClass{
 static boolean myMethod(String input){
  return true;
 }

}

1.

(parameters) -> expression

2.

(parameters) -> { statements; }

Lamda Examples:

() -> {} // valid 1
() -> "Raoul" // valid 1
() -> {return "Mario";} // valid 1
(Integer i) -> return "Alan" + i; // invalid. Valid version:  (Integer i) -> {return "Alan" + i;}
(String s) -> {"Iron Man";} // invalid. Valid version:  (String s) -> {return "Iron Man";} 

More valid examples:

(TYPE name) -> FUNCTION(name)
(String s) -> s.toUpperCase() + " !!!"
(Apple a) -> a.getWeigth() > 150

Compiler can deduce the input type from the functional interface.
So you can omit the type in lambdas.

legal usages of lamdas:

List<Apple> greenApples = filter(inventory, a -> "green".equals(a.getColor())); // a has no type. It becomes an Apple automagically

Methods can be references, to use them as lambdas:

Integer::parseInt // local method on class
Apple::getWeight // static method on class
new Apple()::getWeight // on instance

Supplier<Apple> c1 = Apple::new; // supplier takes no pramas, creates Apple. COnstructor is a supplier
Apple a1 = c1.get(); // produce an apple by calling a suplier

Streams introduce internal collection navigation where you don't need to.
As a contra to extrnal navigation, where you explicitely iterate a collection.

Streams describe computations of Data sets,
whereas collections describe the storage and access.

Operations on Streams are parallelized automagically, when using parallelStream()

inventory.stream().filter...  // not parallelized
inventory.parallelStream().filter... // parallelized

The stream methods are separated into:

 .stream().filter(...).distinct().limit(3)... 

 
.stream()...count();
.stream()...collect(toList());
.stream()....forEach(System.out::println);

Intermesiate methods

terminal methods

All the stream-methods chained one after another are executed in a single, internal iteration.

When you apply a sequence of operations to a stream, these are merely saved up. Only when you apply a terminal operation to a stream is anything actually computed. This has the great advantage when you apply several operations (perhaps a filter and a map followed by a terminal operation reduce) to a stream, then the stream has to be traversed only once instead of for each operation.

THe following code would produce the floowing output:

code

        dishes.stream().filter(a -> {
            System.out.println(a.name);
            return true;
        }).distinct().map(a -> {
            System.out.println(a.type);
            return a.name;
        });

output

pork // name
flesh  // type
pumkin // name
vegetables // type

Name and type come one after another or each element.

This means that steps during parallel execution - methods that can block the parallel execution should not be executed inside the Stream-pipeline.

Assume you have two versions of getting prices from futures:

// directly chaining map methods to create Future, get Result is wrong. Because f.get() will be called after every future creation and block further Future creation (Single Pass)
stream.map(Helper::createFuture).map((Future f) -> g.get()).asList()

// collecting futures into collection first is the right parallel way. When we do the first "get" all Futures already exist and computing
List<Future> futures = stream.map(Helper::toFuture).asList();
List<Integer> prices = futures.asStream().map((Future f) -> g.get())

On the Image the upper Situation is depicted

		// filling the stream with content using "iterate"
		Stream.iterate(0l, i -> i + 1)
				.limit(100);

There are plently of collectors, which are availalbe via java.util.stream.Collectors. Just import all static collector methods and vars to use them by name.

import static java.util.stream.Collectors.*;

Glossar

 List<Dish> dishes = menuStream.collect(toList()); 

 Set<Dish> dishes = menuStream.collect(toSet()); 

 Collection<Dish> dishes = menuStream.collect(toCollection(), ArrayList::new) 

 long howManyDishes = menuStream.collect(counting()); 

 int totalCalories = menuStream.collect(summingInt(Dish::getCalories));  

 double avgCalories = menuStream.collect(averagingInt(Dish::getCalories)); 

 IntSummaryStatistics menuStatistics = menuStream.collect(summarizingInt(Dish::getCalories));  

 String shortMenu = menuStream.map(Dish::getName).collect(joining(", ")); 

 Optional<Dish> lightest = menuStream.collect(minBy(comparingInt(Dish::getCalories))); 

 int totalCalories = menuStream.collect(reducing(0, Dish::getCalories, Integer::sum));  

 int howManyDishes = menuStream.collect(collectingAndThen(toList(), List::size));   

 Map<Dish.Type, List<Dish>> dishesByType = menuStream.collect(groupingBy(Dish::getType)); 

 Map<Boolean, List<Dish>> vegetarianDishes =menuStream.collect(partitioningBy(Dish::isVegetarian)); 

	
		/* COLLECTOR
		 * 
		 * Collector<T,A,R>
		 * 
		 *  T - Type of input item in stream
		 *  A - intermediate Type of Objects, during calculation
		 *  R - Result Type
		 *  
		 *  e.g. Collector which collects Stream of Strings to List 
		 *  T - String
		 *  A - List<String>
		 *  R - List<String>
		 *  
		 *  Supplier<A> supplier() - provides initial value of intermediate Type
		 *  BiConsumer<A,T> accumulator() -  input -> the A value calculated so far. T is the next value in stream
		 *  Function<A,R> finisher() - function is executed after iteration over all elements is done
		 *  BinaryOperator<A> combiner() - function which combined two intermediate Objects. This one makes multithreading possible
		 *  Set<Characteristics> characteristics - some optimization hints. UNORDERED, CONCURRENT, IDENTITY_FINISH
		 *  
		 *  
		 *  supplier - ArrayList::new
		 *  accumulator - (List<String> a, String b) -> {a.add(b)}
		 *  finisher - Functions::identity
		 *  combiner - (List<String> a, List<String> b) -> {a.addAll(b)}
		 *  characteristics - return Collections.unmodifyableSet(EnumSet.of(IDENTITY_FINISH, CONCURRENT ))
		 */

Boxed Values are a performace killer. Using a Stream with primitives - improves performance by factor x6

		long max = 90000000;

// LongStream uses primitives (long)
		start = System.currentTimeMillis();
		summ = LongStream.iterate(0l, i -> i + 1)
		.limit(max)
		.reduce(0l, Long::sum);
		System.out.println("Sequential LongStream Ready filling after: "+ (System.currentTimeMillis()-start) +"ms");
		// 220ms

// Stream generates Boxed Values (Long)
		start = System.currentTimeMillis();
		summ = Stream.iterate(0l, i -> i + 1)
		.limit(max)
		.reduce(0l, Long::sum);
		System.out.println("Sequential Stream Ready filling after: "+ (System.currentTimeMillis()-start) +"ms");
		// 1228 ms

Reducing DIshes to Strings in a parallel way:

		// parallel streams of Dishes, reduced to String, which is a concatenated list of names
		String res = getExampleStream().parallel().reduce("", 
				new BiFunction<String, Dish, String>() { 
			@Override
			public String apply(String t, Dish u) {
				return t+u.name;
			} //accumulator - can convert the Stream content to something else (String)

		}, 
			(a,b) -> {
				return a+b;
			} // combiner - used in the case of using parallel streams
		);

It is possible to implement default methods in interfaces.
So multiple inheritance is some sort of allowed now in java!!!

The default interface methods are marked by keyword default

default void methodName(){...

	public class Pazak implements HasDefaultMethodKu, HasDefaultMethodKuToo{
		
		// have to override method #ku() to explicetly say which interface's method to execute 
		@Override
		public void ku() {
			HasDefaultMethodKu.super.ku();
		}
	}
	
	public interface HasDefaultMethodKu{
		default void ku(){
			System.out.println("Say Ku!");
		}
	}
	
	public interface HasDefaultMethodKuToo{
		default void ku(){
			System.out.println("Say Ku Too!");
		}
	}

Return that instead of null to avoid NPE

• explicitely define where the var may be null

Methods

public class Optionals {

	String name = "Сапожник козоед";
	
	public static void main(String[] args) {
		Optionals optionals = new Optionals();
		System.out.println(optionals.getName().get());
				
		// add defaults
		System.out.println(optionals.getName().orElse("defaultName"));
		System.out.println(optionals.getEmptyName().orElse("defaultName"));
	}
	
	
	Optional<String> getName(){
		return Optional.of(name);
	}
	
	Optional<String> getEmptyName(){
		// explicitly say that this var is nullable
		return Optional.ofNullable(null);
//		return Optional.of(null); //		would produce a NullPointerException
	}

}

Like a future with more functional-style interface.

CompletableFuture<Double> f = CompletableFuture.supplyAsync(() -> calculatePrice());

The occured Exceptions are rethrown on get()

			CompletableFuture<Double> f = ...

			try {
				double calculatedPrice = f.get();
				System.out.println(String.format("Everything was fine. Price calculated: "+calculatedPrice));
				return calculatedPrice;
				
			} catch (Exception e1) {
				System.out.println(String.format("I know - an Exception occured earlier on another thread: %s.", e1.getCause()));
				return (double) 0;
			}

The CompletableFuture can be chained, to execute some code when the future has completed the computation.

CompletableFuture<Integer> c = null;
CompletableFuture<String> c2 = null;

CompletableFuture cc = c.thenApply((Integer i) -> "Done");
/* thenApply((FutureType) -> T) -> CompletableFuture<T>
 * Executed, when previous Future is done.
 * Modifies the value "INSIDE" the completableFuture container to T		 */

/* thenRun(() -> Void) -> CompletableFuture<Void>
 * Executes some code (no params are passed), when previous Future is done		 */

/* thenAccept(FutureType -> Void) -> CompletableFuture<Void>
 * Executed, when the previous Future is done
 * VOID will be inside the Future after this call. To preserve some Type T in FUture - use thenApply()
 */

CompletableFuture ccc = c.thenCompose((Integer inte) -> CompletableFuture.runAsync(() -> System.out.println("Done")));
/* thenCompose(FutureType) -> CompletableFuture) -> CompletableFuture
 * Executes a SECOND CompletableFuture, when the first is done    */

// thenCombine
CompletableFuture<Double> aaa = c.thenCombine(c2, (Integer ii, String ss) -> 2.2);
/* thenCombine(CompletableFuture, Function(FutureType1, FutureType2) -> T) -> CompletableFuture<T>
 * Executes FIRST and SECOND ComputableFutures. Produces the results in a Function which itselfe produces a ComputableFunction 
 */

Chaining can be used in streams, to implement real parallel (not only concurrent) evaluation.

						// List<Picture> -> List<Futures<Picture>> 
						(List<Picture> l) -> {
							List<CompletableFuture<Picture>> listFuturesFIlledPrice = l
									.stream()
									
									// Picture -> Futures<Picture>
									// echo result of future when ready
									.map((Picture picturePriceless) -> {
										CompletableFuture<Picture> picFutureWithPrice;
										
										// ASYNC DELAYED: get price
										picFutureWithPrice = CompletableFutures.getFutureFillingPriceInPictureDelayed(picturePriceless);

										// react immediately after calculation is done
										picFutureWithPrice = picFutureWithPrice.thenApply(CompletableFutures::echoPictureWhenFutureReady); 
										return picFutureWithPrice;
										}
									)
									.collect(Collectors.toList());
							return listFuturesFIlledPrice;
						}

Basic classes

		// INSTANT - to be used by machines
		Instant instant = Instant.ofEpochSecond(1454523011);
		// nanosecond adjustment - 1sec=1.000.000.000nanosec
		Instant.ofEpochSecond(0); 					// 1.January.1970 UTC
		Instant.ofEpochSecond(2, 1000_000_000); 	// 2sec + 1sec
		Instant.ofEpochSecond(3, -1000_000_000); 	// 4sec - 1sec
		
		
		// Instant -> LocalDate, LocalTime
		LocalDate dateOfInstant = LocalDateTime.ofInstant(instant, ZoneId.systemDefault()).toLocalDate();
		LocalTime timeOfInstant = LocalDateTime.ofInstant(instant, ZoneId.systemDefault()).toLocalTime();
		

		LocalDate localDateNow = LocalDate.now();
		LocalDate localDateNowLondon = LocalDate.now(ZoneId.of("UTC+0"));
		LocalDate localDate = LocalDate.of(2016, 2, 16);

		int fieldYear = localDate.get(ChronoField.YEAR);

		LocalTime time = LocalTime.of(14, 55, 54, 11 ); // hour, minute, second, milli
		LocalTime time2 = LocalTime.of(14, 55, 54 ); // hour, minute, second
		LocalTime time3 = LocalTime.of(14, 55 ); // hour, minute
		int hourTime = time.getHour();
		int minuteTime = time.getMinute();

		// 1415-07-6T6:55:20
		LocalDate dateOne = LocalDate.of(1415, Month.JULY, 6);
		LocalTime timeOne = LocalTime.of(6, 55, 20);
		LocalDateTime dt1 = LocalDateTime.of(2014, Month.MARCH, 18, 13, 45, 20);
		LocalDateTime dt2 = LocalDateTime.of(dateOne, timeOne);
		LocalDateTime dt3 = dateOne.atTime(13, 45, 20);
		LocalDateTime dt4 = dateOne.atTime(timeOne);
		LocalDateTime dt5 = timeOne.atDate(dateOne);

		ZoneId romeZone = ZoneId.of("Europe/Rome");
		ZoneId berlinZone = ZoneId.of("Europe/Berlin");
		ZoneId myZone = TimeZone.getDefault().toZoneId();

		ZoneId berlinZone = ZoneId.of("Europe/Berlin");
		
		ZonedDateTime zonedDateTimeBerlin1 = dateOne.atStartOfDay(berlinZone);
		ZonedDateTime zonedDateTimeBerlin2 = localDateTimeOfGuss.atZone(berlinZone);

		LocalDate parsedDate;
		parsedDate = LocalDate.parse("2014-03-18");
		parsedDate = LocalDate.parse("2014-03-18", DateTimeFormatter.ISO_DATE);
		parsedDate = LocalDate.parse("18/03/2014", DateTimeFormatter.ofPattern("dd/MM/yyyy"));
		parsedDate = LocalDate.parse("18.03.2014", DateTimeFormatter.ofPattern("dd.MM.yyyy"));
		
		parsedDate = LocalDate.parse("18 März 2014", DateTimeFormatter.ofPattern("dd MMMM yyyy", Locale.GERMAN));
		
		DateTimeFormatter germanFormatter = new DateTimeFormatterBuilder()
				.appendText(ChronoField.DAY_OF_MONTH)   // 6
				.appendLiteral(".")
				.appendValue(ChronoField.MONTH_OF_YEAR) // 7
				.appendLiteral("( alias ")
				.appendText(ChronoField.MONTH_OF_YEAR) // July
				.appendLiteral(")")
				.appendLiteral(".")
				.appendValue(ChronoField.YEAR)			// 1941
				.parseCaseInsensitive()
				.toFormatter(Locale.GERMAN);
		System.out.println(germanFormatter.format(dateOne));

		/*
		 * Duration is CRAP, because based on Seconds / Nanos.
		 * - you only can compute a duration on "time based units" - something what can handle seconds like LocalTime. 
		 *   The compiler still allows computing Duration between LocalDate, or getting YEARS from Duration,
		 *   which results in an Exception.
		 *   WHY OR WHY CAN'T I JUST CONVERT SECONDS IN ANY UNIT WITHOUT EXCEPTIONS, AS EXPECTED?!
		 */
		// Duration only works with Seconds-supporting types. Using LocaDate - throws an exception
		// Duration durationSinceGuss = Duration.between(dateOne, LocalDate.now());
		Duration durationSinceGuss1 = Duration.between(LocalDateTime.of(dateOne, LocalTime.MIDNIGHT),LocalDateTime.now());
		Duration durationSinceGuss2 = Duration.between(timeOne,LocalTime.now());
//		Duration durationSinceGuss3 = Duration.between(dateOne,LocalDate.now());   		// exception. It trys to convert LocalDate to Seconds, which is not allowed

		Period periodBetweenGuss = Period.between(dateOne, LocalDate.now());
		long yearsSinceGuss7 = periodBetweenGuss.getYears();

LocalDate dateNextSunay = dateOne.with(TemporalAdjusters.next(DayOfWeek.SUNDAY));

The Time Units are contained by ChronoUnit.class and separated in timeBased and dateBased:

It is FUCKING UNCLEAR, why DAYS ARE NOT TIMEBASED?! Still it is important, some operations are only legal for time- or dateBased units: like getting the amount of years from a Duration.

The same separation exists with Classes:

mixing them up produces an UnsupportedTemporalTypeException.

localDate.plus(1, CHronoUnit.SECOND) // UnsupportedTemporalTypeException 
localTime.plus(1, CHronoUnit.DAY)   // UnsupportedTemporalTypeException 

// LocalDateTime supports both
localDateTime.plus(1, ChronoUnit.SECONDS);
localDateTime.plus(1, ChronoUnit.DAYS);

CHronofield - contains diffrent kinds of data info names, which may be requested from diffrent TemporalFields.

		int fieldYear = localDate.get(ChronoField.YEAR);
		/*
                   ChronoField.
		        NanoOfSecond
			NanoOfDay
			MicroOfSecond
			MicroOfDay
			MilliOfSecond
			MilliOfDay
			SecondOfMinute
			SecondOfDay
			MinuteOfHour
			MinuteOfDay
			HourOfAmPm
			ClockHourOfAmPm
			HourOfDay
			ClockHourOfDay
			AmPmOfDay
			DayOfWeek
			AlignedDayOfWeekInMonth
			AlignedDayOfWeekInYear
			DayOfMonth
			DayOfYear
			EpochDay
			AlignedWeekOfMonth
			AlignedWeekOfYear
			MonthOfYear
			ProlepticMonth
			YearOfEra
			Year
			Era
			InstantSeconds
			OffsetSeconds
		 */

How to write code in Java8?

You can replace many anonymous classes with method calls. Important:

Runnable runnable = new Runnable(){
 void execute(){
  System.out.println("Go!")
 }
}

Runnable runnable = () -> System.out.println("Go!"); // ok

int mVar = 20;
Runnable runnable = () -> {
  int mVar = 30; // not ok - overriding vars not allowed
  System.out.println("Go!");
}


public static void doSomething(Runnable r){ r.run(); }
public static void doSomething(Task a){ r.execute(); }
doSomething(() -> System.out.println("Go!")); // not ok -  call doSOmething for Task or Runnable?
doSomething((Task)() -> System.out.println("Go!")); // ok

Stream<String> stream = ...
stream.map(String::length);
stream.map(Helpers::quote);

String quote(String str){
 return String.format("'%s'",str);
}

Native collectors are known by others. Are better readable.

int totalCalories = menu.stream().collect(summingInt(Dish::getCalories));

Better readability, optimization behind the scenes.

// collecting names of dishes, which have > 300 calories
		List<String> dishNames = new ArrayList<>();
		for (Dish dish : menu) {
			if (dish.getCalories() > 300) {
				dishNames.add(dish.getName());
			}
		}
// intent is better readable.
menu.parallelStream()
.filter(d -> d.getCalories() > 300)
.map(Dish::getName)
.collect(toList());

// message is evaluated EVERY TIME, even if the level does not fit
logger.log(Level.FINER, "Problem: " + generateDiagnostic());

// to avoid - pass a lambda, which will be called ONLY IF LEVEL FITS
logger.log(Level.FINER, () -> "Problem: " + generateDiagnostic());

// same code, which is executed around some variable logic 
void method1(){
 doPreparation();
 //variable method here
 varMethod2();
 cleanUp();
}
void method2(){
 doPreparation();
 //variable method here
 varMethod2();
 cleanUp();
}


// better
void method(Runnable r){
 doPreparation();
 r.execute();
 cleanUp();
}
method(()->varMethod1());
method(()->varMethod2());

With Java8 the programming style is pushing towards the functional edge.
Some rules for writing functional code:

Requirenments to functions

Functions may only change private variables.

Calling same function always returns the same result.
E.g. readNextLine() is NOT referential transparent. It alwas returns another line.

What to do, when programming:

pass functions to methods. Return new functions with functionality around it.

 
Function<String, String> transformationPipeline
= addHeader.andThen(Letter::checkSpelling)
.andThen(Letter::addFooter);

create FUnctions producing FUnctions,
which will prefill some constants.

 
// carrying
Function createConvertingFunction(double course){
 return (double x) -> course * x;
}

Function eurToChf = createConvertingFunction(0.9);
Function bintcoinToEur = createConvertingFunction(421);

When function reuses another function and
when preselecting some arguments (but not all) have been passed -
then the function is partially applied

Function echoColoredObject(Object object, String color){
 (object, color) -> System.out.println("The "+object+" has the color "+color);
}

// partially apply function echoColoredObject
Function echoColoredBall(String color){
 (color) -> echoColoredObject("Ball", color)
}

If you would like to create a method accepting a lamda use the Argument, typed with one of those methods: java.util.function.IntBinaryOperator.

Examples:

Consumer<Double> c = aDouble -> System.out.println(aDouble);  // no output

Function<Double, String> f = aDouble -> String.valueOf(f);  // function returns what you define
DoubleFunction<String> d = doubleValue -> String.valueOf(doubleValue); // same as previous

IntBinaryOperator plusOperation = (a, b) -> a + b; // operator on two int numbers a and b

LongPredicate l = longValue -> longValue < 0l; // boolean function

            List<Integer> list = List.of(1, 2, 3, 4, 9, 8, 7, 6, 1, 2, 3);

            // gives maybe 7, 8 or 9, cause findAny doesnt respect order
            int res = list
                    .stream()
                    .parallel()
                    .filter(integer -> integer > 6)
                    .findAny()
                    .orElseThrow();
                    
            assertThat(res, anyOf(is(7), is(8), is(9)));



            // gives 9, cause its the first in the row match to filter
            int res2 = list
                    .stream()
                    .parallel()
                    .filter(integer -> integer > 6)
                    .findFirst()
                    .orElseThrow();

            assertThat(res, is(9));

        // cruel method to concat numbers, just to demonstrate reduce
        List<Integer> list = List.of(1, 2, 3, 4, 9, 8, 7, 6, 1, 2);

        Integer res = list
                .stream()
                .mapToInt(value -> value)
                .reduce(0,
                        (integer, integer2) -> {
                            return Integer.parseInt(String.format("%s%s", integer, integer2));
                        }
                );
        System.out.println(res); // 1234987612

        // better way would be with java8
        String res2 = list
                .stream()
                .map(String::valueOf)
                .collect(Collectors.joining());
        System.out.println(res2); // 1234987612