[pylint] Implement nan-comparison (PLW0117)

[augustelalande]

Summary

Implement pylint's nan-comparison, part of #970.

Test Plan

Text fixture was added.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

ℹ️ ecosystem check detected linter changes. (+48 -0 violations, +0 -0 fixes in 1 projects; 42 projects unchanged)

pandas-dev/pandas (+48 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview

+ pandas/core/arrays/arrow/array.py:1409:29: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/arrays/categorical.py:1597:82: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/arrays/string_arrow.py:710:20: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/base.py:641:34: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/internals/managers.py:1000:48: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/strings/accessor.py:347:63: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/core/strings/object_array.py:102:28: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arithmetic/test_object.py:71:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arithmetic/test_object.py:75:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arrays/categorical/test_analytics.py:107:30: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arrays/categorical/test_analytics.py:117:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arrays/categorical/test_indexing.py:292:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arrays/categorical/test_indexing.py:301:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/arrays/floating/test_contains.py:12:12: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/frame/methods/test_astype.py:762:28: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/frame/methods/test_replace.py:1541:21: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/groupby/test_categorical.py:1496:20: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/categorical/test_indexing.py:345:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/categorical/test_indexing.py:353:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/categorical/test_indexing.py:366:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/categorical/test_indexing.py:376:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/categorical/test_indexing.py:386:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/multi/test_indexing.py:822:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/multi/test_indexing.py:825:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/numeric/test_indexing.py:533:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/period/test_indexing.py:784:16: PLW0117 Comparing against a NaN value; use `math.isnan` instead
+ pandas/tests/indexes/period/test_indexing.py:785:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/indexes/period/test_indexing.py:790:16: PLW0117 Comparing against a NaN value; use `math.isnan` instead
+ pandas/tests/indexes/period/test_indexing.py:791:16: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/io/excel/test_writers.py:1021:38: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/io/excel/test_writers.py:1072:50: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:150:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:151:27: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:152:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:153:27: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:154:27: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:155:23: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:157:29: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:158:30: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:159:29: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:160:30: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:161:30: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/libs/test_libalgos.py:162:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/reductions/test_reductions.py:1362:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/reductions/test_reductions.py:1375:30: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/series/indexing/test_setitem.py:1106:52: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/series/indexing/test_setitem.py:619:23: PLW0117 Comparing against a NaN value; use `np.isnan` instead
+ pandas/tests/series/test_constructors.py:1095:26: PLW0117 Comparing against a NaN value; use `np.isnan` instead

Changes by rule (1 rules affected)

code	total	+ violation	- violation	+ fix	- fix
PLW0117	48	48	0	0	0

[augustelalande]

All the violations in pandas are false positives and are mostly checking for the actual np.nan object i.e. na_value is np.nan, but these seem more like exceptions to the rule, not reasons to change it. np.nan is np.nan will return True, but it's a bad idea to check for a nan value by checking x is np.nan.

Interested to hear people's thoughts.

[augustelalande]

Also interested to hear if there is a better way to do this

[charliermarsh]

I tweaked it a little bit but it's generally correct. Kind of tedious to write out, but correct!

[augustelalande]

Sorry had to fix the typo