SOLR-18304: Optimize collapse performance for String fields in Solr 9.x and later

[bartoszfidrysiak]

Related links:

• https://issues.apache.org/jira/browse/SOLR-18304
• https://lists.apache.org/thread/39jy5rfk4sz9cvcbp9l21pq5ldr1ltng

Description

Migrating to Lucene90DocValuesProducer in Solr 9 revealed a significant performance regression in collapse queries sorted by string fields, due to the extra overhead of LZ4 decompression. Solr 9’s collapse implementation does not apply any optimizations and always calls Lucene’s TermOrdValLeafComparator.copy(), which triggers LZ4.decompress() for every document processed by the collapse query. This decompression overhead did not exist in Solr 8.

SortFieldsCompare.testAndSetGroupValues() / setGroupValues()
  → TermOrdValLeafComparator.copy()
    → BaseSortedDocValues.lookupOrd()
      → TermsDict.seekExact()
        → TermsDict.decompressBlock()
          → LZ4.decompress()   ← bottleneck

Solution

This PR proposes two improvements:

1. Load string-sorted doc values lazily for group heads, materializing the string only when a competing document appears.
2. Avoid loading or materializing string-sorted doc values for documents in the same segment during collapse. Use ordinals instead - they’re numeric, cheaper to compare, and don’t need decompression.

The first improvement focuses on scenarios where many collapse groups contain only a single document, or where collapse sorting uses multiple fields with a string field acting as a tie-breaker.

The second improvement is expected to deliver major gains in cases where many documents originate from the same segment.

Tests

Four new tests were added in TestCollapseQParserPlugin:

• testCollapseStringSortLazyLoadingTieDoesNotEvictGroupHead - verifies that when two documents in the same group have an equal string sort value, the first-seen document remains the group head (a tie must not trigger eviction). Covers both single-segment (ordinal fast path) and multi-segment (slow path) cases.
• testCollapseStringSortOrdinalFastPathMultiClauseTieBreaking - verifies that when clause-1 of a multi-clause sort ties on ordinal comparison, clause-2 correctly decides the winner. Also exercises the remaining-values copy loop with a cross-segment competitor.
• testCollapseStringSortWithoutDocValuesSkipsLazyLoadingAndOrdinalFastPath- verifies that sorting on a string field without SORTED DocValues produces correct results via the eager field-comparator path, ensuring the lazy loading and ordinal fast path are safely bypassed when unavailable.
• testCollapseStringSortOrdinalFastPathDescendingWithMissingValues - verifies that missing values rank last even under a descending sort, where the missing-value sentinel (missingOrd = -1) combined with reverseMul = -1 must still produce the correct ordering in the ordinal fast path.

In addition to TestCollapseQParserPlugin, the CollapsingSearch benchmark was introduced to compare average execution times for collapse queries across different sort field combinations. In the benchmark, documents from the same groups are distributed evenly across all segments. The benchmark was executed locally on my machine against Solr 9.10.1, 10.x branch, main branch, and corresponding snapshot versions, which include the enhancements.

Solr 9.10.1 vs Solr 9.10.1-SNAPSHOT

10.x branch vs 10.x-SNAPSHOT

main branch vs main-SNAPSHOT

Conclusions

• The collapseByDateAndStr benchmark shows that the snapshot (with the enhancements) performs significantly better regardless of the number of segments. This is because the string field serves only as a tiebreaker in the collapse sort, so in most cases comparing dates is sufficient to determine the winner. In addition, string doc values are loaded lazily, which avoids eagerly materializing the string value when it is not needed.
• The collapseByStr benchmark shows that the snapshot (with the enhancements) delivers significantly better performance only when the number of segments is small, especially when most documents from the same group are located in the same segment. In the single-segment case, string doc values do not need to be materialized to pick a winner, since comparing element ordinals is enough and is both safe and efficient. According to the benchmark data, these two improvements combined made collapseByStr much faster for a single segment. With many segments, however, and with documents from the same groups spread evenly across them, the ordinal fast path provides no benefit because most comparisons still require string materialization.

Checklist

Please review the following and check all that apply:

• I have reviewed the guidelines for How to Contribute and my code conforms to the standards described there to the best of my ability.
• I have created a Jira issue and added the issue ID to my pull request title.
• I have given Solr maintainers access to contribute to my PR branch. (optional but recommended, not available for branches on forks living under an organisation)
• I have developed this patch against the main branch.
• I have run ./gradlew check.
• I have added tests for my changes.
• I have added documentation for the Reference Guide
• I have added a changelog entry for my change

[dsmiley]

Great work here!

[Copilot]

Pull request overview

This PR addresses a Solr 9 collapse-query performance regression when collapse sorting includes STRING fields backed by Lucene’s SORTED DocValues (where per-doc lookupOrd() can trigger frequent LZ4 decompression after the Lucene upgrade). It does so by deferring string materialization until needed and by comparing ordinals within the same segment to avoid decompression work.

Changes:

• Teach collapse group-head comparison to store STRING sort values lazily (store (dv, ord) and only lookupOrd() when a competitor actually requires it).
• Add an ordinal “fast path” for STRING sort comparisons when both documents are from the same segment (safe ordinal comparison).
• Add new correctness tests for ties, multi-clause tie-breaking, non-DV STRING sorts, and descending sorts with missing values; add a benchmark and benchmark schema support.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
solr/core/src/java/org/apache/solr/search/CollapsingQParserPlugin.java	Implements lazy STRING sort value handling and same-segment ordinal comparison fast path inside collapse group-head selection.
solr/core/src/java/org/apache/solr/search/LazyStringValue.java	New helper to defer SortedDocValues.lookupOrd() and deep-copy only when materialization is required.
solr/core/src/test/org/apache/solr/search/TestCollapseQParserPlugin.java	Adds new tests covering tie behavior, multi-clause tie-breaking, non-DV STRING sorts, and missing-value handling in descending sorts.
solr/benchmark/src/java/org/apache/solr/bench/search/CollapsingSearch.java	Adds a JMH benchmark to measure collapse performance across different sort patterns/segment counts.
solr/benchmark/src/resources/configs/cloud-minimal/conf/schema.xml	Adds *_s_dv dynamic field to support benchmark string docValues fields.
changelog/unreleased/SOLR-18304-fix-collapse-on-string-performance.yml	Adds changelog entry for SOLR-18304 (currently missing required type value).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[dsmiley]

Going to merge this tonight or maybe tomorrow.
Again -- great work here. The benchmark was above & beyond 🔥

[bartoszfidrysiak]

@dsmiley
It looks like some tests unrelated to my changes are flaky. Could you rerun the build until it passes?

[dsmiley]

I will after the LazyStringValue is moved

[bartoszfidrysiak]

@dsmiley
Could you clarify what you mean by "after the LazyStringValue is moved"? Currently, the class is defined in the same package as CollapsingQParserPlugin (the only caller), which allows it to be package-scoped. I see two options:

• Make it a private static inner class of CollapsingQParserPlugin
• Move it to a different package (if so, which one? Any suggestion?)

I'd like to confirm the intended approach before making the change.

[dsmiley]

Woops... I had pending review feedback but didn't submit! Submitting now.

[dsmiley]

oh one thing... so this is only used by CollapsingQParserPlugin. Shouldn't it be an inner class?
Yes that thing is huge and has many classes ... but still. It's better in there than in the search package which is a big package of lots of stuff.

[bartoszfidrysiak]

Done