With an executor, block grouping's TopGroups.maxScore includes groups that are not returned, so it changes with the index slicing (and with groupOffset > 0 it counts the skipped groups); regression in 10.5.0

Finding LFV-239Area: grouping / block grouping (TopGroups.mergeBlockGroups and BlockGroupingCollectorManager.reduce; GroupingSearch(Query groupEndDocs) on a searcher with an executor)Priority: lowImpact: silent wrong resultsNovelty: new

Finding page for LFV-239

AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.

What it shows

You index each group as a document block and run block grouping (GroupingSearch(Query groupEndDocs)) with a field group sort on an IndexSearcher that has an executor. You expect TopGroups.maxScore to be the same number whether the search runs sequentially or concurrently: sequential block grouping and two-pass term grouping report the best score of the returned groups, and the javadoc says 'Highest score across all hits'. Since 10.5.0 the concurrent value is the maximum over every group that any slice kept, including groups that are not returned, so it changes with the slicing (sequential 1.0, concurrent 5.0 from an unreturned group, best hit 9.0 in the repro), while the returned groups and documents are correct.

The test shows that block grouping (GroupingSearch(Query groupEndDocs), group sort on a numeric field, top 1 group) reports a different TopGroups.maxScore when the searcher has an executor and the default slicing splits 6 segments into 2 slices.

Expected vs actual

Expected: the same maxScore with and without an executor (1.0, the only returned group); actual (-ea and -da): 2.0, the best score of the second slice's top group, which is not returned (mergeBlockGroups takes the maximum of the per-slice maxScore values).

Files

• src/org/apache/lucene/search/grouping/TestBlockGroupingMaxScoreSlicing.java (JUnit test)

Run it

• Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
• Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
• Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
• How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
• Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.

# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble   (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:$LUCENE/lucene/grouping/build/libs/lucene-grouping-$V.jar
CP=$CP:$LUCENE/lucene/queries/build/libs/lucene-queries-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

javac -d classes -cp "$CP" $(find src -name '*.java')

# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
     -cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.search.grouping.TestBlockGroupingMaxScoreSlicing

Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).

Feedback

Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.

Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.