Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Check ahead if we can get the count (#13899) 

Currently, we traverse the BKD tree or perform a binary search using DocValues first, and then check whether the count can be obtained in the count() method of IndexSortSortedNumericDocValuesRangeQuery.

we should consider providing a mechanism to perform this check beforehand, avoid unnecessary processing when dealing with a sparseRange
This commit is contained in:
Lu Xugang 2024-10-25 15:05:51 +08:00 committed by GitHub
parent 2ec5cc8c17
commit 0bbef32cf5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 37 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
lucene/CHANGES.txt
View File

@ -63,6 +63,8 @@ Optimizations
* GITHUB#13941: Optimized computation of top-hits on disjunctive queries with
many clauses. (Adrien Grand)
* GITHUB#13899: Check ahead if we can get the count. (Lu Xugang)
Bug Fixes
---------------------
* GITHUB#13832: Fixed an issue where the DefaultPassageFormatter.format method did not format passages as intended

36
lucene/core/src/java/org/apache/lucene/search/IndexSortSortedNumericDocValuesRangeQuery.java
View File

@ -186,10 +186,44 @@ public class IndexSortSortedNumericDocValuesRangeQuery extends Query {
@Override
public int count(LeafReaderContext context) throws IOException {
if (context.reader().hasDeletions() == false) {
IteratorAndCount itAndCount = getDocIdSetIteratorOrNull(context);
if (lowerValue > upperValue) {
return 0;
}
IteratorAndCount itAndCount = null;
LeafReader reader = context.reader();
// first use bkd optimization if possible
SortedNumericDocValues sortedNumericValues = DocValues.getSortedNumeric(reader, field);
NumericDocValues numericValues = DocValues.unwrapSingleton(sortedNumericValues);
PointValues pointValues = reader.getPointValues(field);
if (pointValues != null && pointValues.getDocCount() == reader.maxDoc()) {
itAndCount = getDocIdSetIteratorOrNullFromBkd(context, numericValues);
}
if (itAndCount != null && itAndCount.count != -1) {
return itAndCount.count;
}
// use index sort optimization if possible
Sort indexSort = reader.getMetaData().sort();
if (indexSort != null
&& indexSort.getSort().length > 0
&& indexSort.getSort()[0].getField().equals(field)) {
final SortField sortField = indexSort.getSort()[0];
final SortField.Type sortFieldType = getSortFieldType(sortField);
// The index sort optimization is only supported for Type.INT and Type.LONG
if (sortFieldType == Type.INT || sortFieldType == Type.LONG) {
Object missingValue = sortField.getMissingValue();
final long missingLongValue = missingValue == null ? 0L : (long) missingValue;
// all documents have docValues or missing value falls outside the range
if ((pointValues != null && pointValues.getDocCount() == reader.maxDoc())
|| (missingLongValue < lowerValue || missingLongValue > upperValue)) {
itAndCount = getDocIdSetIterator(sortField, sortFieldType, context, numericValues);
}
if (itAndCount != null && itAndCount.count != -1) {
return itAndCount.count;
}
}
}
}
return fallbackWeight.count(context);
}