For each element i in each list perform an operation. Elements can be processed in any order. For example in old java:
List<A> aList;
List<B> bList; // aList is larger than bList
for (int i=0; i<bList.size(), i++) {
aList.get(i).doSomethingWith(bList.get(i));
}
for (int j=i; j<aList.size(), j++) {
aList.get(j).doSomething();
}
Which is the best way to implement this with java.util.stream.Stream so elements can be processed in parallel?
You need to work on both lists in parallel so I don't think you can stream the lists themselves. However you can stream the indices and work on that:
IntStream.range(0, aList.size())
.parallel()
.forEach(i -> {
if (i < bList.size()) aList.get(i).doSomethingWith(bList.get(i));
else aList.get(i).doSomething();
});
Just because Stream is new it does not mean that you should forget about all other tools Java provides. Even using these good old tools becomes smoother using Java 8:
List<A> aList;
List<B> bList; // aList is larger than bList
ExecutorService exec = Executors.newCachedThreadPool();
int a=aList.size(), b=bList.size();
assert a>b;
Future<?> f1=exec.submit(()->IntStream.range(0, b)
.parallel().forEach(i->aList.get(i).doSomethingWith(bList.get(i)))
);
Future<?> f2=exec.submit(()->aList.subList(b, a)
.stream().parallel().forEach(A::doSomething)
);
f1.get();
f2.get();
exec.shutdown();
java.util.stream package. I agree your solution makes use of Java standard API, for this reason I upvoted it. However, considering Stream already provides concurrency, I personally believe that using an Executor its redundant, but that's my opinion. Didn't mean to offend you @Holger.