CODEWITHSUNDEEP
javalambdafilterjava-8java-stream
AhmedBakran
AhmedBakran

Reputation: 135

Using streams, lambda in java 8

I would like to optimize the below code to take an existing ArrayList and make a new list based on the below logic. Does streaming/lambda in Java 8 make it possible?

for (ShipmentTracingDTO tracing : tracings) {
    if (tracing.getDestination()) {
        newTracings.add(tracing);
        break;
    }
    newTracings.add(tracing);
}

Upvotes: 8

Views: 264

Answers (3)

Naman
Naman

Reputation: 32046

You can do it using IntStream.range and List.subList possibly as :

List<ShipmentTracingDTO> newTracings = tracings.subList(0, IntStream.range(0, tracings.size())
       .filter(i -> tracings.get(i).getDestination())
       .map(a -> a + 1) // to include the index of occurrence
       .findFirst()
       .orElse(tracings.size())); // else include all until last element

Note: This is just a view of tracings and hence changing the underlying list tracings would also reflect in changes in newTracings. If you want to decouple these, you can create a new ArrayList<>() from the solutions.

Edit: In comments from Holger, you could alternatively use:

List<ShipmentTracingDTO> newTracings = IntStream.rangeClosed(1, tracings.size())
        .filter(i -> tracings.get(i - 1).getDestination()) // i-1 to include this as well
        .mapToObj(i -> tracings.subList(0, i)) // Stream<List<ShipmentTracingDTO>>
        .findFirst() // Optional<List<ShipmentTracingDTO>>
        .orElse(tracings); // or else the entire list

Upvotes: 9

Klitos Kyriacou
Klitos Kyriacou

Reputation: 11664

First, I would point out there's not much wrong with the original code that you want to change. Just because Java has streams, you don't have to use them.

To use streams, I would break up the solution into two phases:

OptionalInt indexOfMatchingItem = IntStream.range(0, tracings.size())
        .filter(i -> tracings.get(i).getDestination())
        .findFirst();

List<ShipmentTracingDTO> newTracings = new ArrayList<>(
        indexOfMatchingItem
                .map(i -> tracings.subList(0, i + 1))
                .orElse(tracings));

The above can be written as a single expression, but splitting it with a suitably named intermediate variable can make the code self-documenting.

I've created a new ArrayList so that it won't be affected by any subsequent changes to the original list tracings. If tracings is immutable, you can skip the construction of a new ArrayList.

The above solution is slightly better performing than the original code in the question, because the ArrayList constructor pre-allocates an array of exactly the required size, and so avoids the overhead of resizing and multiple array copies.

An alternative (perhaps more readable but less performant) solution is:

List<ShipmentTracingDTO> newTracings =
        Stream.concat(
            tracings.stream().takeWhile(i -> !tracings.get(i).getDestination()),
            tracings.stream().dropWhile(i -> !tracings.get(i).getDestination()).limit(1)
        ).collect(toList());

Upvotes: 4

Michał Marcinkowski
Michał Marcinkowski

Reputation: 620

If I needed such processing I would divide it into two separate streams:

final int lastElemPosition = tracings.size() - 1;
final int firstElemWithGetDestinationTruePosition = IntStream.range(0, lastElemPosition)
    .filter(i -> tracings.get(i).getDestination())
    .findFirst()
    .orElse(lastElemPosition);
final List<ShipmentTracingDTO> desiredElements = IntStream.rangeClosed(0, firstElemWithGetDestinationTruePosition)
    .mapToObj(tracings::get)
    .collect(Collectors.toList());

Upvotes: 4

Related Questions