Omega359 opened a new pull request, #26245: URL: https://github.com/apache/datafusion/pull/26245
Depends on #26212; this branch includes its plan-diagnostics commits. ## Which issue does this PR close? No linked issue. ## Rationale for this change Add the existing sort pushdown workloads to the SQL benchmark runner as a standalone suite. Switching `bench.sh` to the SQL runner is deferred to a later PR. ## What changes are included in this PR? Add 30 cases across `baseline`, `sorted`, `inexact`, `inexact_unsorted`, and `inexact_overlap`, with suite metadata, table registration SQL, and usage documentation. The cases reference the existing query files and reuse the existing Parquet datasets. The suite preserves the legacy ordering declarations, statistics collection, and inexact filter-pushdown settings. Data generation remains in the unchanged `bench.sh`; the legacy benchmark remains available. The known result-validation ordering issue is outside this change. ## What is the testing strategy for this PR? Compared the SQL and legacy runners on the same datasets using two rounds of five iterations per case, reversing runner order for the second round (600 timed executions). All row counts matched. The sums of per-case median times differed by -0.6%, -0.1%, +4.4%, -3.9%, and -6.6% across the five subgroups, respectively. Focused rechecks resolved two timing outliers; sorted query 2 remained about 16% faster in the SQL runner at roughly 2 ms. A separate verifier found matching physical plans for all 30 cases and correctly ordered outputs. Full-row fingerprints matched in 29 cases; overlap query 3 differed only in permitted ties at the LIMIT boundary. Original base-data checksums were unchanged. Release builds, documentation formatting, and `git diff --check` passed. ## Are there any user-facing changes? The standalone suite is available from `benchmarks/` with: ```bash cargo run --release --bin benchmark_runner -- sort_pushdown --iterations 5 ``` Use `--subgroup` and `--query` to select a variant and original query number. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
