perf(join): fall back to chained build on pathological key duplication#259
Merged
Conversation
…ntrol, boundary, not-sticky
… fallback on dense moderate dup
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes the O(dup²) build pathology in the radix-parallel join at its source. The per-partition open-addressing hash build is row-granular, so a key with
dupduplicate rows forms adup-length linear-probe run — inserting the k-th scans k slots → quadratic build (a 10M single-key build took ~9.8s). The build now counts same-hash (per-key) duplicates during insert and, when one key exceedsRADIX_DUP_RUN_MAX(512), abandons the radix attempt and re-runs the whole join through the existing chained-HT path (O(1) insert / O(matches) probe — already the trusted reference for small joins and OOM recovery).This is sub-project 4, piece 2. It complements the merged build-side selection (PR #251): that dodges the quadratic for INNER joins by building the smaller side; this fix covers LEFT/FULL (which can't swap) and any forced-build-duplicated-side case. ANTI joins use a separate exec path and are unaffected.
Perf (gate verdict: WIN)
Full data in `bench/bottleneck/join_dup_fallback_compare.md`.
Test Plan