GH-47481: [C++][Acero] record_batch_reader_source specify Ordering::Implicit to support select * limit k

[egolearner]

Rationale for this change

record_batch_reader_source does not support select * limit k as described in #47481

What changes are included in this PR?

record_batch_reader_source specify Ordering::Implicit

Are these changes tested?

yes

Are there any user-facing changes?

no

• GitHub Issue: [C++][Acero] record_batch_reader_source does not support select * limit 3 #47481

[github-actions]

Thanks for opening a pull request!

If this is not a minor PR. Could you open an issue for this pull request on GitHub? https://github.com/apache/arrow/issues/new/choose

Opening GitHub issues ahead of time contributes to the Openness of the Apache Arrow project.

Then could you also rename the pull request title in the following format?

GH-${GITHUB_ISSUE_ID}: [${COMPONENT}] ${SUMMARY}

or

MINOR: [${COMPONENT}] ${SUMMARY}

See also:

• Other pull requests
• Contribution Guidelines - Contributing Overview

[github-actions]

⚠️ GitHub issue #47481 has been automatically assigned in GitHub to PR creator.

[egolearner]

Hi @westonpace , please take a look

[egolearner]

Hi @westonpace , please take a look

[egolearner]

ping @westonpace

[egolearner]

Since @westonpace seems unavailable — @jonkeane @kou @raulcd could you please review this PR?

[egolearner]

ping @westonpace @jonkeane @kou @raulcd

[kou]

Can we add a test for this case?

[Copilot]

Pull request overview

This PR addresses GH-47481 by making record_batch_reader_source report an implicit (meaningful) ordering, allowing Acero’s fetch node (used for LIMIT/OFFSET) to accept it without requiring an explicit order_by.

Changes:

• Mark RecordBatchReaderSourceNode output ordering as Ordering::Implicit() so fetch no longer rejects it as non-deterministic.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[egolearner]

Can we add a test for this case?

UT added. Thx @kou .

[zanmato1984]

+1. I think this change aligns the ordering declaration of record_batch_reader_source with other similar sources (e.g. SchemaSourceNode). Thanks for working on this.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 7c9d880.

There weren't enough matching historic benchmark results to make a call on whether there were regressions.

The full Conbench report has more details.