alexanderbianchi opened a new pull request, #27306: URL: https://github.com/apache/datafusion/pull/27306
## Which issue does this PR close? No linked issue. Adds the execution benchmark requested in #27055. ## Rationale for this change The existing hash-join queries do not isolate propagation of a join-key predicate from a filtered input to the other scan. ## What changes are included in this PR? Add `hj/Q26`: select one order inside a derived table and join it to `lineitem`. Disable dynamic filtering to isolate static predicate inference. Reuse TPC-H data and the existing SQL benchmark runner. No optimizer or `bench.sh` changes. ## What is the testing strategy for this PR? Ran Q26 locally against TPC-H SF1 with the upstream-base optimizer and four partitions. Its data and HashJoinExec assertions pass; the join produces four matches. Formatting and all-targets/all-features Clippy pass. ```sh cargo run --release --bin benchmark_runner -- hj --query 26 --scale-factor 1 --path <data-root> ``` The data root must contain `tpch_sf1/`; use `--scale-factor 10` for `tpch_sf10/`. The optimizer improvement and comparative measurements are in #27055, separately from this benchmark. ## Are there any user-facing changes? None; benchmark coverage only. -- 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]
