RFR: 8362958: Unnecessary copying / sorting in Streams using Comparator.naturalOrder()
Viktor Klang
vklang at openjdk.org
Tue Nov 11 14:39:26 UTC 2025
On Tue, 11 Nov 2025 02:43:14 GMT, Patrick Strawderman <duke at openjdk.org> wrote:
> When `Comparator.naturalOrder()` was explicitly supplied to a collection such as `TreeSet`, or passed into the `sorted` method of a stream, the sorted characteristic was not preserved, causing unnecessary buffering and duplicate sorting.
>
> Example:
>
>
> TreeSet<Integer> sortedSet = new TreeSet<>(Comparator.naturalOrder());
> sortedSet.add(1);
> sortedSet.add(2);
> // SortedOps.OfRef.opWrapSink is not a no-op
> sortedSet.stream().sorted().forEach(System.out::println);
>
> or
>
> TreeSet<Integer> sortedSet = new TreeSet<>();
> sortedSet.add(1);
> sortedSet.add(2);
> // SortedOps.OfRef.opWrapSink is not a no-op
> sortedSet.stream().sorted(Comparator.naturalOrder()).forEach(System.out::println);
>
>
> This PR updates `SortedOps.makeRef` and `StreamOpFlag.fromCharacteristics` to handle the above cases and avoid the unnecessary sort step.
src/java.base/share/classes/java/util/stream/SortedOps.java line 64:
> 62: Comparator<? super T> comparator) {
> 63: return Comparator.naturalOrder().equals(comparator) ?
> 64: new OfRef<>(upstream) : new OfRef<>(upstream, comparator);
I think it would be better to address this directly in the OfRef constructor, and let the 1-arg constructor delegate to the 2-arg constructor.
src/java.base/share/classes/java/util/stream/StreamOpFlag.java line 754:
> 752: if ((characteristics & Spliterator.SORTED) != 0 &&
> 753: !(spliterator.getComparator() == null ||
> 754: spliterator.getComparator().equals(Comparator.naturalOrder()))) {
The following would avoid having to call getComparator() twice and be all &&s.
Suggestion:
if ((characteristics & Spliterator.SORTED) != 0 &&
spliterator.getComparator() instanceof Comparator c &&
!c.equals(Comparator.naturalOrder()))) {
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/28226#discussion_r2514463670
PR Review Comment: https://git.openjdk.org/jdk/pull/28226#discussion_r2514464014
More information about the core-libs-dev
mailing list