Sort.getPrimarySortField (new in 10.6.0, #16424) skips an index-sort field that the reader's FieldInfos do not list, or whose skip index holds a single value, even when that field still orders the documents (a FilterLeafReader hides it but keeps the index sort, or a BinarySortField derives its key from other fields), so doc-values range queries on the next sort field, LongField.newRangeQuery included, return wrong hits or throw

Finding LFV-860Area: search / index-sort range-query shortcuts (Sort.getPrimarySortField; SortedNumericDocValuesRangeQuery, SortedSetDocValuesRangeQuery, IndexSortSortedNumericDocValuesRangeQuery)Priority: mediumImpact: silent wrong resultsNovelty: new

Finding page for LFV-860

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

What it shows

You sort an index by two fields, say (c, f), and then either read it through a FilterLeafReader that hides c from FieldInfos but, like every FilterLeafReader by default, still reports the index sort (c, f) in getMetaData(), or sort by a BinarySortField subclass whose key is derived from another doc-values field, as its javadoc allows; a doc-values range query on f should still match exactly the documents whose f is in the range, as it does on 10.5.2. On main and 10.6.0, Sort.getPrimarySortField (added by #16424) finds no FieldInfo for c (or for the BinarySortField's own name), concludes that sorting by it 'has no effect', and returns f as the primary sort, so the range queries' index-sort shortcuts binary-search a segment that is not ordered by f. The result: most ranges miss matching documents and some return documents outside the range (43 of 55 ranges wrong in an independent test; f:[1 TO 3] returns 0 hits instead of 2), and LongField.newRangeQuery can throw IllegalArgumentException 'minDoc must be < maxDoc'.

The test shows that with an index sort (BinarySortField subclass whose key comes from doc-values field a, then f), Sort.getPrimarySortField skips the key field (no FieldInfo for its name) and returns f, so a doc-values range query on f uses its index-sort shortcut on a segment that is not sorted by f.

Expected vs actual

Expected: count(f:[1 TO 1]) = 1. Actual: expected:<1> but was:<0>.

Files

Run it

# 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:resources:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.search.TestDerivedSortKeyRangeQuery

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.