andygrove opened a new pull request, #1969: URL: https://github.com/apache/datafusion-ballista/pull/1969
# Which issue does this PR close? Closes #1967. # Rationale for this change When viewing distributed plans, `UnresolvedShuffleExec` renders only its partitioning, so it isn't obvious which input stage it reads from: ``` UnresolvedShuffleExec: partitioning: Hash([c_custkey@0], 16) ``` Including the input stage id makes the stage graph legible directly from the plan text (and from `EXPLAIN`/plan-stability snapshots): ``` UnresolvedShuffleExec: stage=3, partitioning: Hash([c_custkey@0], 16) ``` # What changes are included in this PR? - `UnresolvedShuffleExec`'s `Display` (`fmt_as`, Default/Verbose) now prefixes `stage=N` in both the partitioning and broadcast forms, using the existing `stage_id` field. The `TreeRender` form is unchanged. - This is a display-only change — serialization uses the `stage_id` struct field, not the Display string, so plan reconstruction is unaffected. - Updated the affected inline plan-string assertions in `planner.rs` and `client/tests/context_checks.rs`, and regenerated the TPC-H plan-stability goldens (the only change is the added `stage=N` annotation; stage counts are unchanged). # Are there any user-facing changes? Plan text (e.g. `EXPLAIN`, executor logs) now shows the input stage id on `UnresolvedShuffleExec`. No API changes. -- 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]
