Module id.xfunction
Package id.xfunction.function

Class ParallelConsumer<T>

Object
ParallelConsumer<T>
All Implemented Interfaces:
AutoCloseable, Consumer<T>
public class ParallelConsumer<T> extends Object implements Consumer<T>, AutoCloseable
Parallel streams may partition the stream in any order which means that there is no guarantee that items which are in the top of the stream will be processed first. Here is an example: 

 AtomicInteger c = new AtomicInteger();
 Consumer<Integer> consumer = s -> {
     if (c.incrementAndGet() < 15)
         System.out.println(s);
 };

 System.out.println("Parallel stream:");
 range(0, 1000)
     .parallel()
     .boxed()
     .forEach(consumer);
 System.out.println("Total: "  + c);
May produce: 
 Parallel stream:
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 77
 Total: 1000

The output of course differs every time you run the code.

Parallel consumer on the other hand allows you to process items of the stream in parallel but guaranteeing that items will be taken in order. 


 System.out.println("Parallel consumer:");
 c.set(0);
 try (ParallelConsumer<Integer> pconsumer = new ParallelConsumer<>(consumer)) {
     range(0, 1000)
         .boxed()
         .forEach(pconsumer);
 }
 System.out.println("Total: "  + c);
May produce: 
 Parallel consumer:
 0
 1
 4
 5
 2
 3
 6
 7
 8
 9
 10
 12
 13
 11
 Total: 1000

It is achieved by making consumers to take items from the internal queue.

If your application finishes but does not terminate it may be because:

  • you are leaking parallel consumers (make sure to close them)
  • one of your consumers still running preventing JVM from stopping

All exceptions in worker threads by default will be printed to stderr unless the exception handler is defined.