[Ayush-CS-89112521]

Rationale for this change

RoundUtil::Pow10 used float-suffixed constants (1e0F..1e15F), causing double scale factors to truncate to float32 precision and produce off-grid rounded values. In addition, computing fractional parts via round_val - std::floor(round_val) caused catastrophic cancellation on negative numbers near halfway points, triggering false ties.

What changes are included in this PR?

• Replaced float-suffixed constants in Pow10 with type T constants (T(1e0)..T(1e15)).
• Replaced std::floor with std::abs(round_val - std::trunc(round_val)) in RoundToMultiple, Round, and RoundBinary.
• Added regression tests for scale precision, negative halfway ties, and signed zeros.

Are these changes tested?

Yes, added unit tests in scalar_round_arithmetic_test.cc.

Are there any user-facing changes?

Just a bugfix.

Was AI used for this PR?

PR code and description written by:

• Human
• AI

Reviewed before submission by:

• Human
• AI
• Not reviewed

• GitHub Issue: [C++][Compute] Floating round uses float32 scale constants and can return off-grid values #55393

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

[github-actions]

⚠️ GitHub issue #55393 has been automatically assigned in GitHub to PR creator.

[Reranko05]

@Ayush-CS-89112521 Could you follow the Arrow PR template and add AI disclosure checkbox in the PR desc. Thanks!

[Ayush-CS-89112521]

Thanks for pointing that out, @Reranko05! Updated the PR description with the AI disclosure section.

[github-actions]

⚠️ GitHub issue #55393 has been automatically assigned in GitHub to PR creator.