BUG: Boolean type cannot be compared with strigs

[eli-osherovich]

It seems that Pandas' boolean dtype has some bug -- comparison with strings fails.

Code to reproduce (version 1.3.4)

import pandas as pd
s = pd.Series([True, False, pd.NA], dtype="boolean")
s == "X"

The code above fails with this error:

TypeError                                 Traceback (most recent call last)
<ipython-input-41-f42702e3c612> in <module>
----> 1 a == "X"

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/ops/common.py in new_method(self, other)
     67         other = item_from_zerodim(other)
     68 
---> 69         return method(self, other)
     70 
     71     return new_method

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/arraylike.py in __eq__(self, other)
     30     @unpack_zerodim_and_defer("__eq__")
     31     def __eq__(self, other):
---> 32         return self._cmp_method(other, operator.eq)
     33 
     34     @unpack_zerodim_and_defer("__ne__")

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/series.py in _cmp_method(self, other, op)
   5500 
   5501         with np.errstate(all="ignore"):
-> 5502             res_values = ops.comparison_op(lvalues, rvalues, op)
   5503 
   5504         return self._construct_result(res_values, name=res_name)

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/ops/array_ops.py in comparison_op(left, right, op)
    268     ):
    269         # Call the method on lvalues
--> 270         res_values = op(lvalues, rvalues)
    271 
    272     elif is_scalar(rvalues) and isna(rvalues):

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/ops/common.py in new_method(self, other)
     67         other = item_from_zerodim(other)
     68 
---> 69         return method(self, other)
     70 
     71     return new_method

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/arraylike.py in __eq__(self, other)
     30     @unpack_zerodim_and_defer("__eq__")
     31     def __eq__(self, other):
---> 32         return self._cmp_method(other, operator.eq)
     33 
     34     @unpack_zerodim_and_defer("__ne__")

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/arrays/boolean.py in _cmp_method(self, other, op)
    671                 mask = self._mask | mask
    672 
--> 673         return BooleanArray(result, mask, copy=False)
    674 
    675     def _arith_method(self, other, op):

~/virtenvs/mkl/lib/python3.9/site-packages/pandas/core/arrays/boolean.py in __init__(self, values, mask, copy)
    297     def __init__(self, values: np.ndarray, mask: np.ndarray, copy: bool = False):
    298         if not (isinstance(values, np.ndarray) and values.dtype == np.bool_):
--> 299             raise TypeError(
    300                 "values should be boolean numpy array. Use "
    301                 "the 'pd.array' function instead"

TypeError: values should be boolean numpy array. Use the 'pd.array' function instead

[eli-osherovich]

The same bug exists in 1.3.5

[simonjayhawkins]

Thanks @eli-osherovich for the report.

The same bug exists in 1.3.5

was fixed in commit: [0f81a63] BUG: BooleanArray raising on comparison to string (#44533)

the fix will be included in the 1.4 release early in the new year.

import pandas as pd

print(pd.__version__)
s = pd.Series([True, False, pd.NA], dtype="boolean")
s == "X"

0    False
1    False
2     <NA>
dtype: boolean

closing as the PR that fixed added tests explicitly for test_eq_mismatched_type and well as replace.