Skip to content

feat: use range partitioning for date_trunc aggregation - #2

Closed
benbellick wants to merge 4 commits into
fix/issue-25425-date-trunc-scalar-roundingfrom
feat/issue-25345-range-date-trunc-aggregation-v2
Closed

benbellick wants to merge 4 commits into
fix/issue-25425-date-trunc-scalar-roundingfrom
feat/issue-25345-range-date-trunc-aggregation-v2

Conversation

@benbellick

Copy link
Copy Markdown
Owner

Which issue does this PR close?

None. This is a fork-only preview for reviewing the combined change.

Rationale for this change

Grouped aggregation currently inserts a hash repartition when grouping by a monotonic transformation of a range-partitioned key, even when the concrete range boundaries keep every resulting group within one input partition.

What changes are included in this PR?

An argument-type-aware, default-deny audit capability enables only timezone-less date_trunc. Concrete boundary analysis establishes group locality rather than general partition-layout compatibility, so it can be consumed by aggregation without authorizing co-partitioned joins.

Aggregation enforcement and validation share one locality policy across combined and uncombined plans. Non-range arguments must be literals, and unsupported inputs, overlapping buckets, null or extreme boundaries, and evaluation errors fail closed. date_bin, timezone-bearing timestamps, dynamic precision, grouping sets, and joins remain excluded.

This preview is temporarily stacked on the scalar date_trunc correctness fix. That prerequisite will disappear from the optimization diff after it is available on main.

Are these changes tested?

Tests cover all timestamp units and supported granularities, fractional and pre-epoch values, calendar boundaries, nulls, extremes, ascending and descending ranges, compound keys, repeated optimization, uncombined aggregates, and join isolation. A table-driven harness evaluates actual arrays assigned to source ranges and verifies that accepted transformations have disjoint partition images.

The full extended workspace test suite, strict all-target/all-feature Clippy, and the targeted SQLLogicTest pass.

A reproducible Criterion benchmark over aligned hourly partitions measured approximately 35–36% lower execution time on two synthetic workloads by eliminating hash repartitioning.

Are there any user-facing changes?

Eligible grouped aggregations can avoid hash repartitioning. Public range-analysis and aggregate-locality APIs are added.


Note: This PR was developed with AI assistance. All changes have been reviewed, and I take full responsibility for this contribution.

@benbellick
benbellick changed the base branch from main to fix/issue-25425-date-trunc-scalar-rounding September 18, 2026 18:50
@github-actions

Copy link
Copy Markdown

Thank you for opening this pull request!

Reviewer note: cargo-semver-checks reported the current version number is not SemVer-compatible with the changes in this pull request (compared against the base branch).

Details
     Cloning apache/main
    Building datafusion v55.1.0 (current)
       Built [  57.700s] (current)
     Parsing datafusion v55.1.0 (current)
      Parsed [   0.034s] (current)
    Building datafusion v55.1.0 (baseline)
       Built [  56.979s] (baseline)
     Parsing datafusion v55.1.0 (baseline)
      Parsed [   0.034s] (baseline)
    Checking datafusion v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.906s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [ 118.561s] datafusion
    Building datafusion-expr v55.1.0 (current)
       Built [  27.494s] (current)
     Parsing datafusion-expr v55.1.0 (current)
      Parsed [   0.078s] (current)
    Building datafusion-expr v55.1.0 (baseline)
       Built [  27.832s] (baseline)
     Parsing datafusion-expr v55.1.0 (baseline)
      Parsed [   0.078s] (baseline)
    Checking datafusion-expr v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   1.815s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [  58.883s] datafusion-expr
    Building datafusion-expr-common v55.1.0 (current)
       Built [  19.152s] (current)
     Parsing datafusion-expr-common v55.1.0 (current)
      Parsed [   0.019s] (current)
    Building datafusion-expr-common v55.1.0 (baseline)
       Built [  19.009s] (baseline)
     Parsing datafusion-expr-common v55.1.0 (baseline)
      Parsed [   0.020s] (baseline)
    Checking datafusion-expr-common v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.320s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [  39.892s] datafusion-expr-common
    Building datafusion-functions v55.1.0 (current)
       Built [  30.132s] (current)
     Parsing datafusion-functions v55.1.0 (current)
      Parsed [   0.086s] (current)
    Building datafusion-functions v55.1.0 (baseline)
       Built [  30.300s] (baseline)
     Parsing datafusion-functions v55.1.0 (baseline)
      Parsed [   0.085s] (baseline)
    Checking datafusion-functions v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.548s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [  62.701s] datafusion-functions
    Building datafusion-physical-expr v55.1.0 (current)
       Built [  28.050s] (current)
     Parsing datafusion-physical-expr v55.1.0 (current)
      Parsed [   0.049s] (current)
    Building datafusion-physical-expr v55.1.0 (baseline)
       Built [  28.480s] (baseline)
     Parsing datafusion-physical-expr v55.1.0 (baseline)
      Parsed [   0.049s] (baseline)
    Checking datafusion-physical-expr v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.536s] 223 checks: 222 pass, 1 fail, 0 warn, 31 skip

--- failure enum_variant_added: enum variant added on exhaustive enum ---

Description:
A publicly-visible enum without #[non_exhaustive] has a new variant.
        ref: https://doc.rust-lang.org/cargo/reference/semver.html#enum-variant-new
       impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.50.0/src/lints/enum_variant_added.ron

Failed in:
  variant PartitioningSatisfaction:KeyLocal in /home/runner/work/datafusion/datafusion/datafusion/physical-expr/src/partitioning.rs:560

     Summary semver requires new major version: 1 major and 0 minor checks failed
    Finished [  58.316s] datafusion-physical-expr
    Building datafusion-physical-optimizer v55.1.0 (current)
       Built [  39.515s] (current)
     Parsing datafusion-physical-optimizer v55.1.0 (current)
      Parsed [   0.022s] (current)
    Building datafusion-physical-optimizer v55.1.0 (baseline)
       Built [  39.765s] (baseline)
     Parsing datafusion-physical-optimizer v55.1.0 (baseline)
      Parsed [   0.022s] (baseline)
    Checking datafusion-physical-optimizer v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.155s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [  81.058s] datafusion-physical-optimizer
    Building datafusion-physical-plan v55.1.0 (current)
       Built [  36.751s] (current)
     Parsing datafusion-physical-plan v55.1.0 (current)
      Parsed [   0.154s] (current)
    Building datafusion-physical-plan v55.1.0 (baseline)
       Built [  37.347s] (baseline)
     Parsing datafusion-physical-plan v55.1.0 (baseline)
      Parsed [   0.156s] (baseline)
    Checking datafusion-physical-plan v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.967s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [  77.217s] datafusion-physical-plan
    Building datafusion-sqllogictest v55.1.0 (current)
       Built [ 100.719s] (current)
     Parsing datafusion-sqllogictest v55.1.0 (current)
      Parsed [   0.022s] (current)
    Building datafusion-sqllogictest v55.1.0 (baseline)
       Built [  97.977s] (baseline)
     Parsing datafusion-sqllogictest v55.1.0 (baseline)
      Parsed [   0.028s] (baseline)
    Checking datafusion-sqllogictest v55.1.0 -> v55.1.0 (no change; assume patch)
     Checked [   0.131s] 223 checks: 223 pass, 31 skip
     Summary no semver update required
    Finished [ 202.162s] datafusion-sqllogictest

@benbellick

Copy link
Copy Markdown
Owner Author

Superseded by the stacked fork-only previews in #3 and #4.

@benbellick benbellick closed this Sep 18, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant