fix: return null for modulo by zero on pandas-like backends (#4036)

[ioseba]

Fixes #4036.

Description

PR #4011 updated // 0 to return null on pandas-like backends via floordiv_null_on_zero. However, % was still using the raw operator, returning 0 for nullable integer dtypes (e.g. Int64, Int32, UInt32), whereas other backends return null.

This PR:

• Introduces mod_null_on_zero in src/narwhals/_pandas_like/utils.py and routes __mod__ / __rmod__ through it in src/narwhals/_pandas_like/series.py.
• Preserves native nullable integer dtypes while ensuring division/modulo by zero safely returns null (<NA>).
• Adds dedicated unit tests in tests/expr_and_series/division_by_zero_test.py across nullable integer types (Int64, Int32, Int8, UInt32) for both scalar and column operands.

Summary by CodeRabbit

• Bug Fixes
  • Modulo operations now return null results when the divisor is zero, including reflected operations and expressions.
  • Results preserve the input dtype in supported pandas-like backends.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 61faf724-18e2-428f-84ec-16dfa9895b62

---

📥 Commits

Reviewing files that changed from the base of the PR and between 4d6f7b0 and 0297c20.

---

📒 Files selected for processing (3)

• src/narwhals/_pandas_like/series.py
• src/narwhals/_pandas_like/utils.py
• tests/expr_and_series/division_by_zero_test.py

---

Included review availability: This review used your included allowance. Your plan provides up to 4 included reviews per hour; 2 remain after this review.

---

---

📝 Walkthrough

Walkthrough

The pandas-like backend now applies null-on-zero handling to modulo operations. Tests cover scalar and column divisors, reflected operations, expressions, and dtype preservation.

Changes

Pandas-like modulo handling

Layer / File(s)	Summary
Shared null-on-zero helpers src/narwhals/_pandas_like/utils.py	A shared helper applies an operation while masking zero divisors. Floor division delegates to it, and a modulo helper uses it.
Series modulo integration and coverage src/narwhals/_pandas_like/series.py, tests/expr_and_series/division_by_zero_test.py	Both Series modulo operand orders use the modulo helper. Tests cover Series and expression operations, zero and null divisors, and dtype preservation.

Priority: ➖ Normal

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Bug fix · Severity of issue fixed: Medium

Suggested reviewers: fbruzzesi

---

Merge Risk: ⚪ Minimal · up to 0297c

No identified issue blocks merging; run the normal test checks before merge.

Pre-merge checks | 4 | 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 11.76% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 17 functions across 3 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check		The title clearly and concisely identifies the main fix: returning null for modulo-by-zero on pandas-like backends.
Description check		The description explains the issue, the implementation, the affected dtypes, and the added test coverage. It includes a related issue reference but does not complete the template checkboxes for PR typ…
Linked Issues check		Issue #4036 requires null results for modulo by zero on pandas-like nullable integer data, for both operand orders and scalar or zero-valued column divisors. The PR adds _null_on_zero and `mod_null_…
Out of Scope Changes check		The changed source files implement the modulo-by-zero behavior from issue #4036. The added tests verify the requested scalar, column, reflected, mixed-divisor, null-scalar, and dtype cases. No unrelat…

---

• Fix all pre-merge checks with AI

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}