1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
|
/*
* Licensed to Elasticsearch under one or more contributor
* license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright
* ownership. Elasticsearch 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.elasticsearch.join.aggregations;
import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.LeafReaderContext;
import org.apache.lucene.index.SortedSetDocValues;
import org.apache.lucene.search.ConstantScoreScorer;
import org.apache.lucene.search.DocIdSetIterator;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.Weight;
import org.apache.lucene.util.Bits;
import org.elasticsearch.common.ParseField;
import org.elasticsearch.common.lease.Releasables;
import org.elasticsearch.common.lucene.Lucene;
import org.elasticsearch.common.util.LongArray;
import org.elasticsearch.common.util.LongObjectPagedHashMap;
import org.elasticsearch.search.aggregations.Aggregator;
import org.elasticsearch.search.aggregations.AggregatorFactories;
import org.elasticsearch.search.aggregations.InternalAggregation;
import org.elasticsearch.search.aggregations.LeafBucketCollector;
import org.elasticsearch.search.aggregations.bucket.SingleBucketAggregator;
import org.elasticsearch.search.aggregations.pipeline.PipelineAggregator;
import org.elasticsearch.search.aggregations.support.ValuesSource;
import org.elasticsearch.search.internal.SearchContext;
import java.io.IOException;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
// The RecordingPerReaderBucketCollector assumes per segment recording which isn't the case for this
// aggregation, for this reason that collector can't be used
public class ParentToChildrenAggregator extends SingleBucketAggregator {
static final ParseField TYPE_FIELD = new ParseField("type");
private final Weight childFilter;
private final Weight parentFilter;
private final ValuesSource.Bytes.WithOrdinals valuesSource;
// Maybe use PagedGrowableWriter? This will be less wasteful than LongArray,
// but then we don't have the reuse feature of BigArrays.
// Also if we know the highest possible value that a parent agg will create
// then we store multiple values into one slot
private final LongArray parentOrdToBuckets;
// Only pay the extra storage price if the a parentOrd has multiple buckets
// Most of the times a parent doesn't have multiple buckets, since there is
// only one document per parent ord,
// only in the case of terms agg if a parent doc has multiple terms per
// field this is needed:
private final LongObjectPagedHashMap<long[]> parentOrdToOtherBuckets;
private boolean multipleBucketsPerParentOrd = false;
public ParentToChildrenAggregator(String name, AggregatorFactories factories,
SearchContext context, Aggregator parent, Query childFilter,
Query parentFilter, ValuesSource.Bytes.WithOrdinals valuesSource,
long maxOrd, List<PipelineAggregator> pipelineAggregators, Map<String, Object> metaData)
throws IOException {
super(name, factories, context, parent, pipelineAggregators, metaData);
// these two filters are cached in the parser
this.childFilter = context.searcher().createNormalizedWeight(childFilter, false);
this.parentFilter = context.searcher().createNormalizedWeight(parentFilter, false);
this.parentOrdToBuckets = context.bigArrays().newLongArray(maxOrd, false);
this.parentOrdToBuckets.fill(0, maxOrd, -1);
this.parentOrdToOtherBuckets = new LongObjectPagedHashMap<>(context.bigArrays());
this.valuesSource = valuesSource;
}
@Override
public InternalAggregation buildAggregation(long owningBucketOrdinal) throws IOException {
return new InternalChildren(name, bucketDocCount(owningBucketOrdinal),
bucketAggregations(owningBucketOrdinal), pipelineAggregators(), metaData());
}
@Override
public InternalAggregation buildEmptyAggregation() {
return new InternalChildren(name, 0, buildEmptySubAggregations(), pipelineAggregators(),
metaData());
}
@Override
public LeafBucketCollector getLeafCollector(LeafReaderContext ctx,
final LeafBucketCollector sub) throws IOException {
if (valuesSource == null) {
return LeafBucketCollector.NO_OP_COLLECTOR;
}
final SortedSetDocValues globalOrdinals = valuesSource.globalOrdinalsValues(ctx);
Scorer parentScorer = parentFilter.scorer(ctx);
final Bits parentDocs = Lucene.asSequentialAccessBits(ctx.reader().maxDoc(), parentScorer);
return new LeafBucketCollector() {
@Override
public void collect(int docId, long bucket) throws IOException {
if (parentDocs.get(docId) && globalOrdinals.advanceExact(docId)) {
long globalOrdinal = globalOrdinals.nextOrd();
assert globalOrdinals.nextOrd() == SortedSetDocValues.NO_MORE_ORDS;
if (globalOrdinal != -1) {
if (parentOrdToBuckets.get(globalOrdinal) == -1) {
parentOrdToBuckets.set(globalOrdinal, bucket);
} else {
long[] bucketOrds = parentOrdToOtherBuckets.get(globalOrdinal);
if (bucketOrds != null) {
bucketOrds = Arrays.copyOf(bucketOrds, bucketOrds.length + 1);
bucketOrds[bucketOrds.length - 1] = bucket;
parentOrdToOtherBuckets.put(globalOrdinal, bucketOrds);
} else {
parentOrdToOtherBuckets.put(globalOrdinal, new long[] { bucket });
}
multipleBucketsPerParentOrd = true;
}
}
}
}
};
}
@Override
protected void doPostCollection() throws IOException {
IndexReader indexReader = context().searcher().getIndexReader();
for (LeafReaderContext ctx : indexReader.leaves()) {
Scorer childDocsScorer = childFilter.scorer(ctx);
if (childDocsScorer == null) {
continue;
}
DocIdSetIterator childDocsIter = childDocsScorer.iterator();
final LeafBucketCollector sub = collectableSubAggregators.getLeafCollector(ctx);
final SortedSetDocValues globalOrdinals = valuesSource.globalOrdinalsValues(ctx);
// Set the scorer, since we now replay only the child docIds
sub.setScorer(new ConstantScoreScorer(null, 1f, childDocsIter));
final Bits liveDocs = ctx.reader().getLiveDocs();
for (int docId = childDocsIter
.nextDoc(); docId != DocIdSetIterator.NO_MORE_DOCS; docId = childDocsIter
.nextDoc()) {
if (liveDocs != null && liveDocs.get(docId) == false) {
continue;
}
if (globalOrdinals.advanceExact(docId)) {
long globalOrdinal = globalOrdinals.nextOrd();
assert globalOrdinals.nextOrd() == SortedSetDocValues.NO_MORE_ORDS;
long bucketOrd = parentOrdToBuckets.get(globalOrdinal);
if (bucketOrd != -1) {
collectBucket(sub, docId, bucketOrd);
if (multipleBucketsPerParentOrd) {
long[] otherBucketOrds = parentOrdToOtherBuckets.get(globalOrdinal);
if (otherBucketOrds != null) {
for (long otherBucketOrd : otherBucketOrds) {
collectBucket(sub, docId, otherBucketOrd);
}
}
}
}
}
}
}
}
@Override
protected void doClose() {
Releasables.close(parentOrdToBuckets, parentOrdToOtherBuckets);
}
}
|
