python data_error ai_generated true

AssertionError: Arrays are not equal (mismatch 100.0%) x: array([nan]) y: array([nan])

ID: python/numpy-nan-comparison

Also available as: JSON · Markdown · 中文
80%Fix Rate
87%Confidence
0Evidence
2024-06-05First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
3.x active — — —

Root Cause

np.testing.assert_array_equal treats NaN == NaN as False by default. Also affects boolean masks built from NaN comparisons.

generic

中文

np.testing.assert_array_equal 默认将 NaN == NaN 视为 False。也会影响基于 NaN 比较构建的布尔掩码。

Workarounds

  1. 95% success 
    np.testing.assert_array_equal(a, b, strict=False) or use np.testing.assert_allclose(a, b, equal_nan=True)
  2. 90% success 
    mask = (a == b) | (np.isnan(a) & np.isnan(b)); assert mask.all()

Dead Ends

Common approaches that don't work:

  1. 90% fail

    Also returns False for NaN

  2. 90% fail

    NaN != NaN, returns False