[
https://issues.apache.org/jira/browse/FLINK-9491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16531430#comment-16531430
]
ASF GitHub Bot commented on FLINK-9491:
---------------------------------------
Github user StefanRRichter commented on a diff in the pull request:
https://github.com/apache/flink/pull/6228#discussion_r199814578
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/state/heap/HeapPriorityQueue.java
---
@@ -0,0 +1,341 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.runtime.state.heap;
+
+import org.apache.flink.runtime.state.InternalPriorityQueue;
+import org.apache.flink.util.CloseableIterator;
+
+import javax.annotation.Nonnegative;
+import javax.annotation.Nonnull;
+import javax.annotation.Nullable;
+
+import java.util.Arrays;
+import java.util.Collection;
+import java.util.Comparator;
+import java.util.Iterator;
+import java.util.NoSuchElementException;
+
+import static org.apache.flink.util.CollectionUtil.MAX_ARRAY_SIZE;
+
+/**
+ * Basic heap-based priority queue for {@link HeapPriorityQueueElement}
objects. This heap supports fast deletes
+ * because it manages position indexes of the contained {@link
HeapPriorityQueueElement}s. The heap implementation is
+ * a simple binary tree stored inside an array. Element indexes in the
heap array start at 1 instead of 0 to make array
+ * index computations a bit simpler in the hot methods. Object
identification of remove is based on object identity and
+ * not on equals.
+ *
+ * <p>Possible future improvements:
+ * <ul>
+ * <li>We could also implement shrinking for the heap.</li>
+ * </ul>
+ *
+ * @param <T> type of the contained elements.
+ */
+public class HeapPriorityQueue<T extends HeapPriorityQueueElement>
implements InternalPriorityQueue<T> {
+
+ /**
+ * The index of the head element in the array that represents the heap.
+ */
+ private static final int QUEUE_HEAD_INDEX = 1;
+
+ /**
+ * Comparator for the contained elements.
+ */
+ private final Comparator<T> elementComparator;
+
+ /**
+ * The array that represents the heap-organized priority queue.
+ */
+ private T[] queue;
+
+ /**
+ * The current size of the priority queue.
+ */
+ private int size;
+
+ /**
+ * Creates an empty {@link HeapPriorityQueue} with the requested
initial capacity.
+ *
+ * @param elementComparator comparator for the contained elements.
+ * @param minimumCapacity the minimum and initial capacity of this
priority queue.
+ */
+ @SuppressWarnings("unchecked")
+ public HeapPriorityQueue(
+ @Nonnull Comparator<T> elementComparator,
+ @Nonnegative int minimumCapacity) {
+
+ this.elementComparator = elementComparator;
+ this.queue = (T[]) new
HeapPriorityQueueElement[QUEUE_HEAD_INDEX + minimumCapacity];
+ }
+
+ @Override
+ @Nullable
+ public T poll() {
+ return size() > 0 ? removeElementAtIndex(QUEUE_HEAD_INDEX) :
null;
+ }
+
+ @Override
+ @Nullable
+ public T peek() {
+ return size() > 0 ? queue[QUEUE_HEAD_INDEX] : null;
+ }
+
+ /**
+ * Adds the element to add to the heap. This element should not be
managed by any other {@link HeapPriorityQueue}.
+ *
+ * @return <code>true</code> if the operation changed the head element
or if is it unclear if the head element changed.
+ * Only returns <code>false</code> iff the head element was not changed
by this operation.
+ */
+ @Override
+ public boolean add(@Nonnull T toAdd) {
+ return addInternal(toAdd);
+ }
+
+ /**
+ * This remove is based on object identity, not the result of equals.
+ *
+ * @return <code>true</code> if the operation changed the head element
or if is it unclear if the head element changed.
+ * Only returns <code>false</code> iff the head element was not changed
by this operation.
+ */
+ @Override
+ public boolean remove(@Nonnull T toStop) {
+ return removeInternal(toStop);
+ }
+
+ @Override
+ public boolean isEmpty() {
+ return size() == 0;
+ }
+
+ @Override
+ @Nonnegative
+ public int size() {
+ return size;
+ }
+
+ public void clear() {
+ size = 0;
+ Arrays.fill(queue, null);
+ }
+
+ @SuppressWarnings({"unchecked"})
+ @Nonnull
+ public <O> O[] toArray(O[] out) {
--- End diff --
It is aligned with the signature from `Collection::toArray(...)`
> Implement timer data structure based on RocksDB
> -----------------------------------------------
>
> Key: FLINK-9491
> URL: https://issues.apache.org/jira/browse/FLINK-9491
> Project: Flink
> Issue Type: Sub-task
> Components: State Backends, Checkpointing
> Reporter: Stefan Richter
> Assignee: Stefan Richter
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.6.0
>
>
> We can now implement timer state that is stored in RocksDB for users that run
> the {{RocksDBKeyedStateBackend}}. As explained in the design document
> (https://docs.google.com/document/d/1XbhJRbig5c5Ftd77d0mKND1bePyTC26Pz04EvxdA7Jc/edit#heading=h.17v0k3363r6q)
> this should also give us asynchronous and incremental snapshots for timer
> state that is larger than main memory.
> We need to think about a way in which to user can select either to run timers
> on RocksDB or on the heap when using the {{RocksDBKeyedStateBackend}}.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
