Github user StefanRRichter commented on a diff in the pull request: https://github.com/apache/flink/pull/4722#discussion_r140821446 --- Diff: flink-runtime/src/main/java/org/apache/flink/runtime/state/heap/CopyOnWriteStateTable.java --- @@ -287,6 +289,14 @@ public S get(K key, N namespace) { } @Override + public Stream<K> getKeys(N namespace) { + return Arrays.stream(primaryTable) --- End diff -- This hash table applies incremental rehashing. This method must also consider the data stored ``incrementalRehashTable`` if it isn't ``null`` or we can miss some data.
---