[Aggregations] Optimize singleton handling in GlobalOrdinalValuesSource (#17740)

* added singleton optimization path to globalordinalvaluesource

Signed-off-by: shreyah963 <[email protected]>

* enabled remote debugging

Signed-off-by: shreyah963 <[email protected]>

* Removed the minimum cap to handle larger ordinal values

Signed-off-by: shreyah963 <[email protected]>

* emove redundant singleton optimization state from SingleDimensionValuesSource and simplify the optimization logic in GlobalOrdinalValuesSource. The singleton optimization is now only applied when DocValues.unwrapSingleton() succeeds, preventing array index out of bounds errors with high cardinality fields.

Signed-off-by: shreyah963 <[email protected]>

* removed redundant initialization

Signed-off-by: shreyah963 <[email protected]>

* reverted the array allocation in the constructer to its original form

Signed-off-by: shreyah963 <[email protected]>

* [Docs] Add detailed comments to GlobalOrdinalValuesSource collector

Signed-off-by: shreyah963 <[email protected]>

* Remote redundant imports and disable remote debugging

Signed-off-by: shreyah963 <[email protected]>

* replaced wildcard import with only necessary imports

Signed-off-by: shreyah963 <[email protected]>

* Update CHANGELOG.md

Signed-off-by: shreyah963 <[email protected]>

* Update CHANGELOG.md

Co-authored-by: bowenlan-amzn <[email protected]>
Signed-off-by: shreyah963 <[email protected]>

* Remove redundant comments from GlobalOrdinalValuesSource

Signed-off-by: shreyah963 <[email protected]>

---------

Signed-off-by: shreyah963 <[email protected]>
Signed-off-by: Ankit Jain <[email protected]>
Co-authored-by: bowenlan-amzn <[email protected]>
Co-authored-by: Ankit Jain <[email protected]>

Author: 3 people

CHANGELOG.md

@@ -31,6 +31,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 ### Changed
 - Migrate BC libs to their FIPS counterparts ([#14912](https://github.com/opensearch-project/OpenSearch/pull/14912))
 - Increase the floor segment size to 16MB ([#17699](https://github.com/opensearch-project/OpenSearch/pull/17699))
+- Unwrap singleton DocValues in global ordinal value source of composite histogram aggregation ([#17740](https://github.com/opensearch-project/OpenSearch/pull/17740))
 - Unwrap singleton DocValues in date histogram aggregation. ([#17643](https://github.com/opensearch-project/OpenSearch/pull/17643))
 - Introduce 512 byte limit to search and ingest pipeline IDs ([#17786](https://github.com/opensearch-project/OpenSearch/pull/17786))
 

server/src/main/java/org/opensearch/search/aggregations/bucket/composite/GlobalOrdinalValuesSource.java

@@ -32,8 +32,10 @@
 
 package org.opensearch.search.aggregations.bucket.composite;
 
+import org.apache.lucene.index.DocValues;
 import org.apache.lucene.index.IndexReader;
 import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.index.SortedDocValues;
 import org.apache.lucene.index.SortedSetDocValues;
 import org.apache.lucene.search.MatchAllDocsQuery;
 import org.apache.lucene.search.Query;
@@ -171,6 +173,26 @@ LeafBucketCollector getLeafCollector(LeafReaderContext context, LeafBucketCollec
         if (lookup == null) {
             initLookup(dvs);
         }
+
+        // unwrapSingleton() returns non-null only if the field is single-valued
+        final SortedDocValues singleton = DocValues.unwrapSingleton(dvs);
+
+        // Direct ordinal access for single-valued fields
+        if (singleton != null) {
+            return new LeafBucketCollector() {
+                @Override
+                public void collect(int doc, long bucket) throws IOException {
+                    if (singleton.advanceExact(doc)) {
+                        currentValue = singleton.ordValue();
+                        next.collect(doc, bucket);
+                    } else if (missingBucket) {
+                        currentValue = -1;
+                        next.collect(doc, bucket);
+                    }
+                }
+            };
+        }
+
         return new LeafBucketCollector() {
             @Override
             public void collect(int doc, long bucket) throws IOException {