API/BUG: pd.NA == pd.NaT returns False

[jbrockmendel]

@jorisvandenbossche I expected this to return pd.NA. Is this intentional?

[jorisvandenbossche]

The same is true for any datetime-like scalar, not just NaT:

In [9]: pd.NA == pd.Timestamp("2012")
Out[9]: False

I don't think this is intentional, but somewhat "knowingly ignored for now", at least going from this comment: #30245 (comment) (although that specific comment was about ufuncs, but for scalar operations it's the same logic).

Right now, we only handle a few specific scalar types (and for other scalars, it defers by returning NotImplemented):

pandas/pandas/_libs/missing.pyx

Lines 287 to 299 in 44e3c40

	def _create_binary_propagating_op(name, is_divmod=False):
	def method(self, other):
	if (other is C_NA or isinstance(other, str)
	or isinstance(other, (numbers.Number, np.bool_))
	or isinstance(other, np.ndarray) and not other.shape):
	# Need the other.shape clause to handle NumPy scalars,
	# since we do a setitem on `out` below, which
	# won't work for NumPy scalars.
	if is_divmod:
	return NA, NA
	else:
	return NA

So either we need to handle more scalar types here, or our scalar types need to handle NA.

[jbrockmendel]

i think the relevant Timestamp/NaT methods have been updated so they will return NotImplemented when operating against pd.NA, so it would now be easy to change the behavior if desired