[vibe bench] feat: implement LeftSingle join type for scalar subquery decorrelation#20999
Closed
Dandandan wants to merge 3 commits intoapache:mainfrom
Closed
[vibe bench] feat: implement LeftSingle join type for scalar subquery decorrelation#20999Dandandan wants to merge 3 commits intoapache:mainfrom
Dandandan wants to merge 3 commits intoapache:mainfrom
Conversation
Introduce a new LeftSingle join operator that behaves like a left outer join but errors at runtime if more than one right-side row matches any given left-side row. This enforces the SQL scalar subquery invariant (at most one row) at the join level rather than relying on recursive evaluation, enabling O(n) hash-based execution. The scalar_subquery_to_join optimizer rule now emits LeftSingle joins instead of Left joins, and the eliminate_outer_join rule can still convert LeftSingle to Inner when the right side is proven non-nullable. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Dandandan
changed the title
Contributor
Author
|
run benchmarks |
- Replace HashSet with bitmap in check_single_match for efficiency - Use DataFusionError::Execution instead of Internal (user-facing error) - Simplify supports_swap to just `!matches!(self, LeftSingle)` - Don't eliminate LeftSingle to Inner (would lose at-most-one constraint) - Use nested match instead of if/else chain in functional_dependencies Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpch — base (merge-base)
tpch — branch
|
Contributor
Author
|
run benchmarks |
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
|
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
|
github-actions
bot
added
sql
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpch — base (merge-base)
tpch — branch
|
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
|
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
|
…gression The eliminate_outer_join optimization is safe for LeftSingle because: - The WHERE clause already filters out NULL (unmatched) rows - The GROUP BY added during scalar subquery decorrelation guarantees at-most-one row per group, so the single-match constraint holds - Without this, LeftSingle is treated like Left for optimization, preventing filter pushdown and causing massive plan regressions Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Contributor
Author
|
run benchmarks |
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpch — base (merge-base)
tpch — branch
|
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
|
|
🤖 Benchmark completed (GKE) | trigger Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
|
Contributor
Author
|
I think this is the main change |
Contributor
|
"[vibe bench]" 😆 |
Contributor
Author
|
Probably not worth a new join type at this point (even though it is not that complex one). |
alamb
reviewed
Mar 18, 2026
Contributor
alamb
left a comment
alamb
left a comment
There was a problem hiding this comment.
Somehow I forgot to submit this review
| /// - If a left row has exactly one match on the right: returns the matched pair | ||
| /// - If a left row has no match on the right: returns the left row with NULLs for right columns | ||
| /// - If a left row has more than one match on the right: returns a runtime error | ||
| LeftSingle, |
Contributor
There was a problem hiding this comment.
At vertica we called this type of join PK join or a Unique join as it assuemes any particular join key is unique
Maybe we can use a similar term here (rather than Single)
LeftUnique 🤔
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
3 participants
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.
Introduce a new LeftSingle join operator that behaves like a left outer join but errors at runtime if more than one right-side row matches any given left-side row. This enforces the SQL scalar subquery invariant (at most one row) at the join level rather than relying on recursive evaluation, enabling O(n) hash-based execution.
The scalar_subquery_to_join optimizer rule now emits LeftSingle joins instead of Left joins, and the eliminate_outer_join rule can still convert LeftSingle to Inner when the right side is proven non-nullable.
Which issue does this PR close?
Rationale for this change
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?