Numeric score joins (JoinUtil numeric overload, PointInSetIncludingScoreQuery) compare equal when only their per-value join scores differ, so DisjunctionMaxQuery keeps only one of two such joins and BooleanQuery merges them, returning wrong, clause-order-dependent scores

Finding LFV-403Area: join / JoinUtil numeric score join (PointInSetIncludingScoreQuery equals/hashCode)Priority: lowImpact: silent wrong resultsNovelty: known upstream

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

What it shows

You build two numeric score joins with JoinUtil.createJoinQuery("fromId", false, "toId", Integer.class, fromQuery, searcher, ScoreMode.Max), one over from-index A and one over from-index B; both collect the join value 7 and score the to-document 0.3648 and 0.4311, and you expect DisjunctionMaxQuery([qa, qb], 0) to give the maximum 0.4311, as the string-key joins over the same data do. Instead qa.equals(qb) is true with equal hash codes, because PointInSetIncludingScoreQuery's equals/hashCode compare the score mode, to field, from query and collected join values but not the per-value scores (aggregatedJoinScores) or multipleValuesPerDocument, so DisjunctionMaxQuery keeps one of the two equal disjuncts (it rewrites to '(toId:{7} | toId:{7})') and scores 0.3648. The same happens with one from searcher and two from fields whose value sets coincide, where the result then depends on clause order, and BooleanQuery SHOULD(qa) SHOULD(qb) is folded into one boosted clause that loses the second join's scores.

Expected vs actual

• Expected: You expect DisjunctionMaxQuery([qa, qb], 0) to give the maximum 0.4311, as the string-key joins over the same data do.
• Actual: qa.equals(qb) is true with equal hash codes, because PointInSetIncludingScoreQuery's equals/hashCode compare the score mode, to field, from query and collected join values but not the per-value scores (aggregatedJoinScores) or multipleValuesPerDocument, so DisjunctionMaxQuery keeps one of the two equal disjuncts (it rewrites to '(toId:{7} | toId:{7})') and scores 0.3648.

Files

• src/org/apache/lucene/search/join/TestNumericScoreJoinEquality.java (JUnit test)

Run it

• Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
• Tested on Lucene main ea57360912ed (2026-10-24): reproduced.
• 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/join/build/libs/lucene-join-$V.jar
CP=$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.join.TestNumericScoreJoinEquality

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 2026-11-03. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.