With an INT index sort, merges order documents by the full 64-bit doc value while flush, sorted search and CheckIndex use its low 32 bits, so merging a value outside the int range writes a segment that violates its declared index sort (CheckIndex at level 2 or higher rejects it) and sorted searches that terminate early on the index sort return wrong top hits (a regression in 8.2.0)
Finding LFV-779Area: index / index sorting at merge time (IndexSorter.IntSorter merge keys, MultiSorter) -> search / TopFieldCollector index-sort early terminationPriority: lowImpact: data corruptionNovelty: new
AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.
What it shows
You configure an index sort on SortField("s", SortField.Type.INT), index documents with NumericDocValuesField("s", value), one of them with a value that does not fit in an int (for example 2^32, whose int value is 0), and let segments merge, expecting the merged segment to be sorted the way a flushed segment of the same documents is. Flush (IntSorter#getDocComparator), the search-time IntComparator and CheckIndex all compare (int) of the value, but the merge key (IntSorter#getComparableProviders) is the full long, so forceMerge(1) of the segments [a = 2^32] and [b = 1, c = 2] writes the order b, c, a instead of a, b, c. CheckIndex (level 2 or higher) then reports 'segment has indexSort=<int: "s"> but docID=1 sorts after docID=2' and a top-1 search sorted by the index sort returns b instead of a, while before the merge the same index is clean and returns a.
The test shows that under an INT index sort, merging segments [a = 2^32] and [b = 1, c = 2] orders documents by the full 64-bit value, while flush and the search-time comparator use the low 32 bits ((int) 2^32 = 0).
Expected vs actual
Expected the merged docID order a, b, c; actual b, c, a (expected:<[abc]> but was:<[bca]>), with -ea and -da.
Files
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.index.TestIntIndexSortMergeOrderGradle 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.