Skip to content

chore: update sqllogictest priority list with latest timing summary (8s --> 6s)#22549

Merged
alamb merged 1 commit into
apache:mainfrom
alamb:worktree-update-slt-timing-order
May 27, 2026
Merged

chore: update sqllogictest priority list with latest timing summary (8s --> 6s)#22549
alamb merged 1 commit into
apache:mainfrom
alamb:worktree-update-slt-timing-order

Conversation

@alamb
Copy link
Copy Markdown
Contributor

@alamb alamb commented May 26, 2026

Which issue does this PR close?

Rationale for this change

Make running sqllogictests faster by improving scheduling / parallelism

The sqllogictest runner runs test files in parallel but each file
sequentially, so it prioritizes known long-running files to run first to
minimize total wall-clock time. This list was last set in #20656.

The relative timings have since shifted. Running with the
--timing-summary flag now reports a different ordering of the longest
files, with nested_loop_join_spill.slt having become the longest by a
fair margin:

$ cargo test --profile=ci --test sqllogictests -- --timing-summary
Per-file elapsed summary (deterministic):
  1.    5.437s  nested_loop_join_spill.slt
  2.    3.471s  push_down_filter_regression.slt
  3.    3.458s  aggregate.slt
  4.    3.065s  joins.slt
  5.    2.852s  aggregate_skip_partial.slt
  6.    2.832s  imdb.slt
  7.    2.453s  window.slt
  8.    1.831s  group_by.slt
  9.    1.282s  clickbench.slt
 10.    1.055s  datetime/timestamps.slt
 11.    0.994s  array/array_has.slt
 12.    0.840s  cte.slt
 13.    0.748s  sort_pushdown.slt
 14.    0.714s  push_down_filter_parquet.slt
 15.    0.668s  projection_pushdown.slt

What changes are included in this PR?

  • Update TEST_PRIORITY_ENTRIES to match the latest --timing-summary ordering
  • Update the example output in the accompanying doc comment.

Are these changes tested?

There are some existing unit tests and I verified timings manually

cargo test --profile=ci --test sqllogictests

Main

Running with 16 test threads (available parallelism: 16)
Completed 475 test files in 8 seconds

This Branch

andrewlamb@Andrews-MacBook-Pro-3:~/Software/datafusion3$ cargo test --profile=ci --test sqllogictests
    Finished `ci` profile [unoptimized] target(s) in 0.53s
     Running bin/sqllogictests.rs (target/ci/deps/sqllogictests-0a94fbc565e8c132)
Running with 16 test threads (available parallelism: 16)
Completed 475 test files in 6 seconds

Are there any user-facing changes?

No.

@github-actions github-actions Bot added optimizer Optimizer rules core Core DataFusion crate sqllogictest SQL Logic Tests (.slt) labels May 26, 2026
@alamb alamb added the development-process Related to development process of DataFusion label May 26, 2026
Refreshes the `TEST_PRIORITY_ENTRIES` list (and its example output) so the
longest-running test files run first, based on the latest `--timing-summary`
output. `nested_loop_join_spill.slt` is now the longest-running file and is
promoted to the front of the list; several additional long-running files
(`array/array_has.slt`, `cte.slt`, `sort_pushdown.slt`,
`push_down_filter_parquet.slt`, `projection_pushdown.slt`) are added.

Follows up on the prioritization mechanism added in apache#20656.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@alamb alamb force-pushed the worktree-update-slt-timing-order branch from ab596b9 to 7434d88 Compare May 26, 2026 20:35
@github-actions github-actions Bot removed development-process Related to development process of DataFusion optimizer Optimizer rules core Core DataFusion crate labels May 26, 2026
@alamb alamb changed the title chore: update sqllogictest priority list with latest timing summary chore: update sqllogictest priority list with latest timing summary (8s --> 6s) May 26, 2026
@alamb alamb marked this pull request as ready for review May 26, 2026 20:47
@alamb alamb added the development-process Related to development process of DataFusion label May 26, 2026
Copy link
Copy Markdown
Contributor

@2010YOUY01 2010YOUY01 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thank you!

@alamb alamb added this pull request to the merge queue May 27, 2026
@alamb
Copy link
Copy Markdown
Contributor Author

alamb commented May 27, 2026

Thank you for the review @2010YOUY01

Merged via the queue into apache:main with commit 786d56f May 27, 2026
38 of 39 checks passed
@alamb alamb deleted the worktree-update-slt-timing-order branch May 27, 2026 18:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

development-process Related to development process of DataFusion sqllogictest SQL Logic Tests (.slt)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants