LateInteractionFloatValuesSource#isCacheable always returns true although it reads the multi-vectors from binary doc values, so after an in-place update of a LateInteractionField the query cache keeps returning the old matches of a FunctionMatchQuery or DoubleRange query over it

Finding LFV-336Area: core / search: LateInteractionFloatValuesSource (late-interaction multi-vectors) and the query cachePriority: lowImpact: silent wrong resultsNovelty: new

Finding page for LFV-336

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

What it shows

You score documents against a query multi-vector with LateInteractionFloatValuesSource, filter on that score through a query cache (for example FunctionMatchQuery(source, s -> s > 0.75), or DoubleRange#getQuery in the facet module), then replace one document's LateInteractionField multi-vector in place with IndexWriter#updateBinaryDocValue and reopen the reader. You expect the reopened searcher to stop matching that document, because a source that reads doc values must report itself not cacheable once those doc values have been updated (SegmentCacheable). Instead isCacheable always returns true, the segment keeps its core cache key, and the cache keeps serving the pre-update result (5 hits instead of 4 in the repro) until the entry is evicted or the segment is merged away.

The test shows that LateInteractionFloatValuesSource#isCacheable still returns true after the LateInteractionField it reads (binary doc values) is updated in place with IndexWriter#updateDocValues, so a query cache can keep serving pre-update matches.

Expected vs actual

Expected false (SegmentCacheable: doc-values readers return DocValues#isCacheable; updated fields are not cacheable); actual: true (expected:<false> but was:<true>), with -ea and -da.

Files

• src/org/apache/lucene/search/TestLateInteractionValuesSourceCacheable.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=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$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.TestLateInteractionValuesSourceCacheable

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).

Formal models and scripts this finding's part of the hunt used: query-cache-correctness

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.