CachingCollector and CachingCollectorManager report isCached() true (no overflow) but hold only the hits their wrapped first-pass collector let through (TOP_SCORES pruning, early leaf termination, skipped leaves), so replay() silently feeds the second pass fewer matches than re-running the query
Finding LFV-161Area: search / CachingCollector first-pass caching (CachingCollector, CachingCollectorManager)Priority: mediumImpact: 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 run a query once through a CachingCollector or CachingCollectorManager that wraps a first-pass collector, see isCached() return true (the cache did not overflow), and replay the cached hits into a second collector, expecting the same result as running that collector on the query, which is what the javadocs and the grouping docs offer ('without re-running the query', 'hold all hits'). Instead the cache holds only the hits the first-pass collector let through: TopScoreDocCollectorManager(10, 1000), the IndexSearcher#search(Query, int) default, prunes non-competitive hits (even when the score-caching variant tells the searcher it needs COMPLETE), an index-order TopFieldCollector stops each segment early, and TotalHitCountCollector skips whole segments it can count from index statistics. In the repro, replaying into a hit counter sees 1001, 2 and 0 of 5000 matching documents, with no error.
Expected vs actual
- Expected: The test passes: it asserts the documented behaviour.
- Actual: The cache holds only the hits the first-pass collector let through: TopScoreDocCollectorManager(10, 1000), the IndexSearcher#search(Query, int) default, prunes non-competitive hits (even when the score-caching variant tells the searcher it needs COMPLETE), an index-order TopFieldCollector stops each segment early, and TotalHitCountCollector skips whole segments it can count from index statistics.
Files
src/org/apache/lucene/search/TestCachingCollectorManagerIncompleteCache.java(JUnit test)src/org/apache/lucene/search/TestCachingCollectorIncompleteCacheLegacy.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=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar # test-framework dependencies (Maven Central)
CP=$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.TestCachingCollectorManagerIncompleteCache
# run 2: assertions on, seed DEADBEEF, default flags
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.TestCachingCollectorIncompleteCacheLegacy
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.