GH-38358: [R] NotImplemented: Function 'if_else' has no kernel matching input types (bool, bool, date32[day])

[thisisnic]

Rationale for this change

Error message which occurred for if_else() when an NA was incorrectly mapped to a logical one, not the right type

What changes are included in this PR?

cast NA to desired type

Are these changes tested?

Yep

Are there any user-facing changes?

Yep

• GitHub Issue: [R] NotImplemented: Function 'if_else' has no kernel matching input types (bool, bool, date32[day]) #38358

[github-actions]

⚠️ GitHub issue #38358 has been automatically assigned in GitHub to PR creator.

[Copilot]

🟡 Changes recommended

The new cast_or_parse() null handling assumes x$is_valid exists (Scalar-only) and can error for Array inputs (used elsewhere), so the guard must be narrowed to Scalars.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Fixes Arrow R dplyr::if_else() failures when a bare NA (logical null scalar) participates in type resolution against date/timestamp columns, by ensuring the NA is represented as a null scalar of the target Arrow type (GH-38358).

Changes:

• Add a special-case in cast_or_parse() to avoid invalid scalar casts (e.g., bool -> date32) by producing a typed null instead.
• Add a regression test covering if_else() with bare NA against Date and POSIXct columns.

File summaries

File	Description
r/R/dplyr-funcs-simple.R	Adjusts scalar casting logic so null scalars don’t attempt unsupported casts when aligning to a common type.
r/tests/testthat/test-dplyr-funcs-conditional.R	Adds a test reproducing the if_else() bare-NA vs date/timestamp failure and validating the fix.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

🟢 Approval recommended

The functional change is small, targeted, and covered by a regression test; the only feedback is a minor comment-clarity improvement.

Review details

Suppressed comments (1)

Previously missed (1) — in code that hasn't changed since the last review.

r/R/dplyr-funcs-simple.R:194

• The comment here says we “skip the value cast and just create a null of the target type”, but the implementation actually creates a null scalar and then casts it. Rewording the comment would make the intent clearer (i.e., we avoid casting the invalid scalar’s inferred type such as bool -> date32).

• Files reviewed: 2/2 changed files
• Comments generated: 0 new
• Review effort level: Lite

[jonkeane]

its the NAs here that are the issue?

[thisisnic]

Yeah, we're making sure they don't get converted to the default logical NAs instead of type-specific NAs

[jonkeane]

One small question, but if I'm not totally confused about that, this is good to merge imo