You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@lucene.apache.org by gs...@apache.org on 2022/01/10 18:44:27 UTC
[lucene] branch branch_9x updated: LUCENE-10356: Further optimize facet counting for single-valued TaxonomyFacetCounts (#585)
This is an automated email from the ASF dual-hosted git repository.
gsmiller pushed a commit to branch branch_9x
in repository https://gitbox.apache.org/repos/asf/lucene.git
The following commit(s) were added to refs/heads/branch_9x by this push:
new 674eb5a LUCENE-10356: Further optimize facet counting for single-valued TaxonomyFacetCounts (#585)
674eb5a is described below
commit 674eb5a4907b420a758a8b8a98f1d0b0722840a7
Author: Greg Miller <gs...@gmail.com>
AuthorDate: Mon Jan 10 10:23:46 2022 -0800
LUCENE-10356: Further optimize facet counting for single-valued TaxonomyFacetCounts (#585)
---
lucene/CHANGES.txt | 2 +
.../facet/taxonomy/FastTaxonomyFacetCounts.java | 68 ++++++++++++++--------
2 files changed, 46 insertions(+), 24 deletions(-)
diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index 9373db11..17006fb 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -112,6 +112,8 @@ Optimizations
* LUCENE-10350: Avoid some duplicate null check in facet counting for TaxonomyFacetCounts. (Guo Feng)
+* LUCENE-10356: Further optimize facet counting for single-valued TaxonomyFacetCounts. (Greg Miller)
+
Changes in runtime behavior
---------------------
diff --git a/lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java b/lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java
index 8975759..fe82726 100644
--- a/lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java
+++ b/lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java
@@ -72,25 +72,40 @@ public class FastTaxonomyFacetCounts extends IntTaxonomyFacets {
private void count(List<MatchingDocs> matchingDocs) throws IOException {
for (MatchingDocs hits : matchingDocs) {
- SortedNumericDocValues dv =
+ SortedNumericDocValues multiValued =
FacetUtils.loadOrdinalValues(hits.context.reader(), indexFieldName);
- if (dv == null) {
+ if (multiValued == null) {
continue;
}
+ NumericDocValues singleValued = DocValues.unwrapSingleton(multiValued);
+
+ DocIdSetIterator valuesIt = singleValued != null ? singleValued : multiValued;
DocIdSetIterator it =
- ConjunctionUtils.intersectIterators(Arrays.asList(hits.bits.iterator(), dv));
+ ConjunctionUtils.intersectIterators(Arrays.asList(hits.bits.iterator(), valuesIt));
- if (values != null) {
- while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
- for (int i = 0; i < dv.docValueCount(); i++) {
- values[(int) dv.nextValue()]++;
+ if (singleValued != null) {
+ if (values != null) {
+ while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ values[(int) singleValued.longValue()]++;
+ }
+ } else {
+ while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ sparseValues.addTo((int) singleValued.longValue(), 1);
}
}
} else {
- while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
- for (int i = 0; i < dv.docValueCount(); i++) {
- sparseValues.addTo((int) dv.nextValue(), 1);
+ if (values != null) {
+ while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ for (int i = 0; i < multiValued.docValueCount(); i++) {
+ values[(int) multiValued.nextValue()]++;
+ }
+ }
+ } else {
+ while (it.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ for (int i = 0; i < multiValued.docValueCount(); i++) {
+ sparseValues.addTo((int) multiValued.nextValue(), 1);
+ }
}
}
}
@@ -102,40 +117,45 @@ public class FastTaxonomyFacetCounts extends IntTaxonomyFacets {
private final void countAll(IndexReader reader) throws IOException {
assert values != null;
for (LeafReaderContext context : reader.leaves()) {
- SortedNumericDocValues dv = FacetUtils.loadOrdinalValues(context.reader(), indexFieldName);
- if (dv == null) {
+ SortedNumericDocValues multiValued =
+ FacetUtils.loadOrdinalValues(context.reader(), indexFieldName);
+ if (multiValued == null) {
continue;
}
Bits liveDocs = context.reader().getLiveDocs();
- NumericDocValues ndv = DocValues.unwrapSingleton(dv);
+ NumericDocValues singleValued = DocValues.unwrapSingleton(multiValued);
- if (ndv != null) {
+ if (singleValued != null) {
if (liveDocs == null) {
- while (ndv.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
- values[(int) ndv.longValue()]++;
+ while (singleValued.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ values[(int) singleValued.longValue()]++;
}
} else {
- for (int doc = ndv.nextDoc(); doc != DocIdSetIterator.NO_MORE_DOCS; doc = ndv.nextDoc()) {
+ for (int doc = singleValued.nextDoc();
+ doc != DocIdSetIterator.NO_MORE_DOCS;
+ doc = singleValued.nextDoc()) {
if (liveDocs.get(doc)) {
- values[(int) ndv.longValue()]++;
+ values[(int) singleValued.longValue()]++;
}
}
}
} else {
if (liveDocs == null) {
- while (dv.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
- final int dvCount = dv.docValueCount();
+ while (multiValued.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ final int dvCount = multiValued.docValueCount();
for (int i = 0; i < dvCount; i++) {
- values[(int) dv.nextValue()]++;
+ values[(int) multiValued.nextValue()]++;
}
}
} else {
- for (int doc = dv.nextDoc(); doc != DocIdSetIterator.NO_MORE_DOCS; doc = dv.nextDoc()) {
+ for (int doc = multiValued.nextDoc();
+ doc != DocIdSetIterator.NO_MORE_DOCS;
+ doc = multiValued.nextDoc()) {
if (liveDocs.get(doc)) {
- final int dvCount = dv.docValueCount();
+ final int dvCount = multiValued.docValueCount();
for (int i = 0; i < dvCount; i++) {
- values[(int) dv.nextValue()]++;
+ values[(int) multiValued.nextValue()]++;
}
}
}