Fixed inefficient Stream API call chains ending with count() (opensea…

…rch-project#15386) (opensearch-project#16361) * Fixed inefficient Stream API call chains ending with count() * Refactored method minTermLength() as per @sandeshkr419's advice * Added a line in CHANGELOG.md --------- Signed-off-by: Dmitry Kryukov <[email protected]> Signed-off-by: Daniel Widdis <[email protected]> Co-authored-by: Dmitry Kryukov <[email protected]>
dbwiddis · Oct 22, 2024 · 0f110e8 · 0f110e8
1 parent 89ed273
commit 0f110e8
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 7 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -88,6 +88,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Fix multi-search with template doesn't return status code ([#16265](https://github.com/opensearch-project/OpenSearch/pull/16265))
 - [Streaming Indexing] Fix intermittent 'The bulk request must be terminated by a newline [\n]' failures [#16337](https://github.com/opensearch-project/OpenSearch/pull/16337))
 - Fix wrong default value when setting `index.number_of_routing_shards` to null on index creation ([#16331](https://github.com/opensearch-project/OpenSearch/pull/16331))
+- Fix inefficient Stream API call chains ending with count() ([#15386](https://github.com/opensearch-project/OpenSearch/pull/15386))
 - [Workload Management] Make query groups persistent across process restarts [#16370](https://github.com/opensearch-project/OpenSearch/pull/16370)
 - Fix typo super->sb in method toString() of RemoteStoreNodeAttribute ([#15362](https://github.com/opensearch-project/OpenSearch/pull/15362))
 - Fix array hashCode calculation in ResyncReplicationRequest ([#16378](https://github.com/opensearch-project/OpenSearch/pull/16378))

diff --git a/modules/percolator/src/main/java/org/opensearch/percolator/QueryAnalyzer.java b/modules/percolator/src/main/java/org/opensearch/percolator/QueryAnalyzer.java
@@ -510,19 +510,25 @@ static Result selectBestResult(Result result1, Result result2) {
     }
 
     private static int minTermLength(Set<QueryExtraction> extractions) {
-        // In case there are only range extractions, then we return Integer.MIN_VALUE,
-        // so that selectBestExtraction(...) we are likely to prefer the extractions that contains at least a single extraction
-        if (extractions.stream().filter(queryExtraction -> queryExtraction.term != null).count() == 0
-            && extractions.stream().filter(queryExtraction -> queryExtraction.range != null).count() > 0) {
-            return Integer.MIN_VALUE;
-        }
-
+        boolean hasTerm = false;
+        boolean hasRange = false;
         int min = Integer.MAX_VALUE;
+
         for (QueryExtraction qt : extractions) {
             if (qt.term != null) {
+                hasTerm = true;
                 min = Math.min(min, qt.bytes().length);
             }
+            if (qt.range != null) {
+                hasRange = true;
+            }
         }
+
+        // If there are no terms but there are ranges, return Integer.MIN_VALUE
+        if (!hasTerm && hasRange) {
+            return Integer.MIN_VALUE;
+        }
+
         return min;
     }