-
Notifications
You must be signed in to change notification settings - Fork 25.6k
Add hash for N integers #135745
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
Merged
Add hash for N integers #135745
Changes from all commits
Commits
Show all changes
2 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
163 changes: 163 additions & 0 deletions
163
server/src/main/java/org/elasticsearch/common/util/IntNHash.java
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,163 @@ | ||
| /* | ||
| * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one | ||
| * or more contributor license agreements. Licensed under the "Elastic License | ||
| * 2.0", the "GNU Affero General Public License v3.0 only", and the "Server Side | ||
| * Public License v 1"; you may not use this file except in compliance with, at | ||
| * your election, the "Elastic License 2.0", the "GNU Affero General Public | ||
| * License v3.0 only", or the "Server Side Public License, v 1". | ||
| */ | ||
|
|
||
| package org.elasticsearch.common.util; | ||
|
|
||
| import com.carrotsearch.hppc.BitMixer; | ||
|
|
||
| import org.elasticsearch.core.Releasables; | ||
|
|
||
| /** | ||
| * Specialized hash table implementation maps N int values to ids. | ||
| * Collisions are resolved with open addressing and | ||
| * linear probing, growth is smooth thanks to {@link BigArrays} and capacity | ||
| * is always a multiple of N for faster identification of buckets. | ||
| * This class is not thread-safe. | ||
| */ | ||
| // IDs are internally stored as id + 1 so that 0 encodes for an empty slot | ||
| public final class IntNHash extends AbstractHash { | ||
| private IntArray keyArray; | ||
| private final int keySize; | ||
| private final int[] scratch; | ||
|
|
||
| // Constructor with configurable capacity and default maximum load factor. | ||
| public IntNHash(long capacity, int keySize, BigArrays bigArrays) { | ||
| this(capacity, keySize, DEFAULT_MAX_LOAD_FACTOR, bigArrays); | ||
| } | ||
|
|
||
| // Constructor with configurable capacity and load factor. | ||
| public IntNHash(long capacity, int keySize, float maxLoadFactor, BigArrays bigArrays) { | ||
| super(capacity, maxLoadFactor, bigArrays); | ||
| this.keySize = keySize; | ||
| this.scratch = new int[keySize]; | ||
| try { | ||
| // `super` allocates a big array so we have to `close` if we fail here or we'll leak it. | ||
| keyArray = bigArrays.newIntArray(keySize * capacity, false); | ||
| } finally { | ||
| if (keyArray == null) { | ||
| close(); | ||
| } | ||
| } | ||
| } | ||
|
|
||
| public int[] getKeys(long id) { | ||
| getKeys(id, scratch); | ||
| return scratch; | ||
| } | ||
|
|
||
| public void getKeys(long id, int[] dst) { | ||
| assert dst.length == keySize; | ||
| for (int i = 0; i < keySize; i++) { | ||
| dst[i] = keyArray.get(keySize * id + i); | ||
| } | ||
| } | ||
|
|
||
| private boolean keyEquals(long id, int[] keys) { | ||
| long keyOffset = keySize * id; | ||
| // TODO: fast equals in BigArray | ||
| for (int i = 0; i < keys.length; i++) { | ||
| if (keyArray.get(keyOffset + i) != keys[i]) { | ||
| return false; | ||
| } | ||
| } | ||
| return true; | ||
| } | ||
|
|
||
| public long find(int[] keys) { | ||
| final long slot = slot(hash(keys), mask); | ||
| for (long index = slot;; index = nextSlot(index, mask)) { | ||
| final long id = id(index); | ||
| if (id == -1) { | ||
| return id; | ||
| } else if (keyEquals(id, keys)) { | ||
| return id; | ||
| } | ||
| } | ||
| } | ||
|
|
||
| private long set(long id, int[] keys) { | ||
| assert size < maxSize; | ||
| long slot = slot(hash(keys), mask); | ||
| for (long index = slot;; index = nextSlot(index, mask)) { | ||
| final long curId = id(index); | ||
| if (curId == -1) { // means unset | ||
| setId(index, id); | ||
| append(id, keys); | ||
| ++size; | ||
| return id; | ||
| } else { | ||
| if (keyEquals(curId, keys)) { | ||
| return -1 - curId; | ||
| } | ||
| } | ||
| } | ||
| } | ||
|
|
||
| private void append(long id, int[] keys) { | ||
| final long keyOffset = keySize * id; | ||
| keyArray = bigArrays.grow(keyArray, keyOffset + keySize); | ||
| for (int i = 0; i < keys.length; i++) { | ||
| keyArray.set(keyOffset + i, keys[i]); | ||
| } | ||
| } | ||
|
|
||
| private void reset(long id) { | ||
| final long slot = slot(hashFromKeyArray(id), mask); | ||
| for (long index = slot;; index = nextSlot(index, mask)) { | ||
| final long curId = id(index); | ||
| if (curId == -1) { // means unset | ||
| setId(index, id); | ||
| break; | ||
| } | ||
| } | ||
| } | ||
|
|
||
| /** | ||
| * Try to add {@code key}. Return its newly allocated id if it wasn't in | ||
| * the hash table yet, or {@code -1-id} if it was already present in | ||
| * the hash table. | ||
| */ | ||
| public long add(int[] keys) { | ||
| if (size >= maxSize) { | ||
| assert size == maxSize; | ||
| grow(); | ||
| } | ||
| assert size < maxSize; | ||
| return set(size, keys); | ||
| } | ||
|
|
||
| @Override | ||
| protected void removeAndAdd(long index) { | ||
| final long id = getAndSetId(index, -1); | ||
| assert id >= 0; | ||
| reset(id); | ||
| } | ||
|
|
||
| @Override | ||
| public void close() { | ||
| Releasables.close(keyArray, super::close); | ||
| } | ||
|
|
||
| static long hash(int[] keys) { | ||
| long hash = BitMixer.mix(keys[0]); | ||
| for (int i = 1; i < keys.length; i++) { | ||
| hash = 31L * hash + BitMixer.mix(keys[i]); | ||
| } | ||
| return hash; | ||
| } | ||
|
|
||
| long hashFromKeyArray(long id) { | ||
| final long keyOffset = id * keySize; | ||
| long hash = BitMixer.mix(keyArray.get(keyOffset)); | ||
| for (int i = 1; i < keySize; i++) { | ||
| hash = 31L * hash + BitMixer.mix(keyArray.get(keyOffset + i)); | ||
| } | ||
| return hash; | ||
| } | ||
| } | ||
190 changes: 190 additions & 0 deletions
190
server/src/test/java/org/elasticsearch/common/util/IntNHashTests.java
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,190 @@ | ||
| /* | ||
| * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one | ||
| * or more contributor license agreements. Licensed under the "Elastic License | ||
| * 2.0", the "GNU Affero General Public License v3.0 only", and the "Server Side | ||
| * Public License v 1"; you may not use this file except in compliance with, at | ||
| * your election, the "Elastic License 2.0", the "GNU Affero General Public | ||
| * License v3.0 only", or the "Server Side Public License, v 1". | ||
| */ | ||
|
|
||
| package org.elasticsearch.common.util; | ||
|
|
||
| import org.elasticsearch.common.settings.Settings; | ||
| import org.elasticsearch.core.Releasables; | ||
| import org.elasticsearch.indices.breaker.NoneCircuitBreakerService; | ||
| import org.elasticsearch.test.ESTestCase; | ||
|
|
||
| import java.util.Arrays; | ||
| import java.util.HashMap; | ||
| import java.util.Map; | ||
|
|
||
| import static org.hamcrest.Matchers.equalTo; | ||
|
|
||
| public class IntNHashTests extends ESTestCase { | ||
| private BigArrays randombigArrays() { | ||
| return new MockBigArrays(new MockPageCacheRecycler(Settings.EMPTY), new NoneCircuitBreakerService()); | ||
| } | ||
|
|
||
| public void test1Key() { | ||
| LongHash hash1 = new LongHash(between(1, 100), randombigArrays()); | ||
| IntNHash hash2 = new IntNHash(randomIntBetween(0, 100), 1, randombigArrays()); | ||
| IntNHash hash3 = new IntNHash(randomIntBetween(0, 100), 2, randombigArrays()); | ||
| int padding = randomInt(); | ||
| int values = between(10, 1000); | ||
| for (int n = 0; n < values; n++) { | ||
| int value = randomIntBetween(0, 1000); | ||
| long id1 = hash1.add(value); | ||
| long id2 = hash2.add(new int[] { value }); | ||
| long id3 = hash3.add(new int[] { value, padding }); | ||
| assertThat(id1, equalTo(id2)); | ||
| assertThat(id1, equalTo(id3)); | ||
| } | ||
| assertThat(hash1.size(), equalTo(hash2.size)); | ||
| assertThat(hash2.size(), equalTo(hash3.size)); | ||
| for (long l = 0; l < hash1.size(); l++) { | ||
| int v1 = (int) hash1.get(l); | ||
| int v2 = hash2.getKeys(l)[0]; | ||
| int v3 = hash3.getKeys(l)[0]; | ||
| assertThat(v1, equalTo(v2)); | ||
| assertThat(v1, equalTo(v3)); | ||
| } | ||
| Releasables.close(hash1, hash2, hash3); | ||
| } | ||
|
|
||
| public void test2Keys() { | ||
| LongHash hash1 = new LongHash(between(1, 100), randombigArrays()); | ||
| IntNHash hash2 = new IntNHash(randomIntBetween(0, 100), 2, randombigArrays()); | ||
| IntNHash hash3 = new IntNHash(randomIntBetween(0, 100), 3, randombigArrays()); | ||
| int padding = randomInt(); | ||
| int values = between(10, 1000); | ||
| for (int n = 0; n < values; n++) { | ||
| int first = randomIntBetween(0, 1000); | ||
| int second = randomIntBetween(0, 1000); | ||
| long v = (((long) first) << 32) | (second & 0xFFFFFFFFL); | ||
| long id1 = hash1.add(v); | ||
| long id2 = hash2.add(new int[] { first, second }); | ||
| long id3 = hash3.add(new int[] { first, second, padding }); | ||
| assertThat(id1, equalTo(id2)); | ||
| assertThat(id1, equalTo(id3)); | ||
| } | ||
| assertThat(hash1.size(), equalTo(hash2.size)); | ||
| assertThat(hash2.size(), equalTo(hash3.size())); | ||
| for (long l = 0; l < hash1.size(); l++) { | ||
| long v1 = hash1.get(l); | ||
| int first = (int) (v1 >>> 32); | ||
| int second = (int) (v1 & 0xFFFFFFFFL); | ||
| int[] v2 = hash2.getKeys(l); | ||
| assertThat(v2, equalTo(new int[] { first, second })); | ||
| int[] v3 = hash3.getKeys(l); | ||
| assertThat(v3, equalTo(new int[] { first, second, padding })); | ||
| } | ||
| Releasables.close(hash1, hash2, hash3); | ||
| } | ||
|
|
||
| public void test3Keys() { | ||
| Int3Hash hash1 = new Int3Hash(between(1, 100), randombigArrays()); | ||
| IntNHash hash2 = new IntNHash(randomIntBetween(0, 100), 3, randombigArrays()); | ||
| IntNHash hash3 = new IntNHash(randomIntBetween(0, 100), 4, randombigArrays()); | ||
| int values = between(10, 1000); | ||
| int padding = randomInt(); | ||
| for (int n = 0; n < values; n++) { | ||
| int v1 = randomIntBetween(0, 1000); | ||
| int v2 = randomIntBetween(0, 1000); | ||
| int v3 = randomIntBetween(0, 1000); | ||
| long id1 = hash1.add(v1, v2, v3); | ||
| long id2 = hash2.add(new int[] { v1, v2, v3 }); | ||
| long id3 = hash3.add(new int[] { v1, v2, v3, padding }); | ||
| assertThat(id1, equalTo(id2)); | ||
| assertThat(id1, equalTo(id3)); | ||
| } | ||
| assertThat(hash1.size(), equalTo(hash2.size)); | ||
| assertThat(hash2.size(), equalTo(hash3.size())); | ||
| for (long l = 0; l < hash1.size; l++) { | ||
| int v1 = hash1.getKey1(l); | ||
| int v2 = hash1.getKey2(l); | ||
| int v3 = hash1.getKey3(l); | ||
| assertThat(hash2.getKeys(l), equalTo(new int[] { v1, v2, v3 })); | ||
| assertThat(hash3.getKeys(l), equalTo(new int[] { v1, v2, v3, padding })); | ||
| } | ||
| Releasables.close(hash1, hash2, hash3); | ||
| } | ||
|
|
||
| public void test4Keys() { | ||
| LongLongHash hash1 = new LongLongHash(between(1, 100), randombigArrays()); | ||
| IntNHash hash2 = new IntNHash(randomIntBetween(0, 100), 4, randombigArrays()); | ||
| IntNHash hash3 = new IntNHash(randomIntBetween(0, 100), 5, randombigArrays()); | ||
| int padding = randomInt(); | ||
| int values = between(10, 1000); | ||
| for (int n = 0; n < values; n++) { | ||
| int v1 = randomIntBetween(0, 1000); | ||
| int v2 = randomIntBetween(0, 1000); | ||
| int v3 = randomIntBetween(0, 1000); | ||
| int v4 = randomIntBetween(0, 1000); | ||
| long id1 = hash1.add((((long) v1) << 32) | (v2 & 0xFFFFFFFFL), (((long) v3) << 32) | (v4 & 0xFFFFFFFFL)); | ||
| long id2 = hash2.add(new int[] { v1, v2, v3, v4 }); | ||
| long id3 = hash3.add(new int[] { v1, v2, v3, v4, padding }); | ||
| assertThat(id1, equalTo(id2)); | ||
| assertThat(id1, equalTo(id3)); | ||
| } | ||
| assertThat(hash1.size(), equalTo(hash2.size)); | ||
| assertThat(hash1.size(), equalTo(hash3.size)); | ||
| for (long l = 0; l < hash1.size; l++) { | ||
| long k1 = hash1.getKey1(l); | ||
| long k2 = hash1.getKey2(l); | ||
| int v1 = (int) (k1 >>> 32); | ||
| int v2 = (int) (k1 & 0xFFFFFFFFL); | ||
| int v3 = (int) (k2 >>> 32); | ||
| int v4 = (int) (k2 & 0xFFFFFFFFL); | ||
| assertThat(hash2.getKeys(l), equalTo(new int[] { v1, v2, v3, v4 })); | ||
| assertThat(hash3.getKeys(l), equalTo(new int[] { v1, v2, v3, v4, padding })); | ||
| } | ||
| Releasables.close(hash1, hash2, hash3); | ||
| } | ||
|
|
||
| public void testLargeKeys() { | ||
| record Ints(int[] vs) { | ||
| @Override | ||
| public boolean equals(Object o) { | ||
| if (this == o) return true; | ||
| if (o == null || getClass() != o.getClass()) return false; | ||
| Ints key = (Ints) o; | ||
| return Arrays.equals(vs, key.vs); | ||
| } | ||
|
|
||
| @Override | ||
| public int hashCode() { | ||
| return Arrays.hashCode(vs); | ||
| } | ||
| } | ||
| Map<Ints, Long> maps = new HashMap<>(); | ||
| int keySize = randomIntBetween(1, 50); | ||
| int values = between(1, 1000); | ||
| try (IntNHash hash = new IntNHash(randomIntBetween(0, 1000), keySize, randombigArrays())) { | ||
| for (int i = 0; i < values; i++) { | ||
| int[] keys = new int[keySize]; | ||
| for (int k = 0; k < keySize; k++) { | ||
| keys[k] = randomIntBetween(0, 1000); | ||
| } | ||
| long hashId = hash.add(keys); | ||
| Ints ints = new Ints(keys); | ||
| if (hashId < 0) { | ||
| hashId = -1 - hashId; | ||
| assertThat(maps.get(ints), equalTo(hashId)); | ||
| } else { | ||
| assertNull(maps.get(ints)); | ||
| maps.put(ints, hashId); | ||
| } | ||
| assertThat((int) hash.size, equalTo(maps.size())); | ||
| } | ||
| for (long l = 0; l < hash.size; l++) { | ||
| int[] key = hash.getKeys(l); | ||
| Ints ints = new Ints(key); | ||
| assertThat(maps.get(ints), equalTo(l)); | ||
| } | ||
| for (var e : maps.entrySet()) { | ||
| long id = hash.find(e.getKey().vs); | ||
| assertThat(id, equalTo(e.getValue())); | ||
| } | ||
| } | ||
| } | ||
| } |
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't we be unsetting the first match that is not -1? I may misunderstood this.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good point. I copy-pasted this from other classes. I think the issue is the method name reset. This method is called after unsetting (in removeAndAdd) and is used to re-add to an empty slot instead.