Rich-T-kid opened a new pull request, #27444: URL: https://github.com/apache/datafusion/pull/27444
## Which issue does this PR close? - Validates apache/arrow-rs#12260, which implements apache/arrow-rs#12195. ## Rationale for this change Run DataFusion's main hosted benchmark suites against the Parquet dictionary-transition optimization in apache/arrow-rs#12260. ## What changes are included in this PR? The first commit pins the Arrow crates used by DataFusion to the Arrow PR parent `2553a2be695299b6c18876cbe511cb642fbe5963`. The second commit changes those pins to the exact PR head `82f7c48b93731b8467fa85255b3dde21d11a192e`. There are no DataFusion runtime changes. This is a benchmark-only PR and must not be merged. ## What is the testing strategy for this PR? The exact PR-head pin passes `cargo fmt --all -- --check` and a locked `release-nonlto` build of `benchmark_runner`. The full pre-push Clippy hook reaches an existing test-only unused import in `datafusion/physical-expr-adapter/src/schema_rewriter.rs`; this dependency-only branch does not change that file. The hosted benchmark bot compares the two reachable DataFusion commits directly with RLE dictionary reads enabled and a 32 GB runtime memory limit. TPC-H, TPC-DS, and partitioned ClickBench results will be linked here when complete. ## Are there any user-facing changes? No. This PR only provides a reachable dependency pin for benchmark validation. ## AI usage disclosure Codex helped prepare and validate the dependency pins and benchmark configuration. -- 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]
