====================================================================== FAILURES =======================================================================
__________________________________________________________ test_col_unicode_sandwich_unicode __________________________________________________________
numpy.core._exceptions._UFuncNoLoopError: ufunc 'not_equal' did not contain a loop with signature matching types (<class 'numpy.dtype[str_]'>, <class 'numpy.dtype[bytes_]'>) -> None
The above exception was the direct cause of the following exception:
def test_col_unicode_sandwich_unicode():
"""
Sanity check that Unicode Column behaves normally.
"""
uba = 'bä'
uba8 = uba.encode('utf-8')
c = table.Column([uba, 'def'], dtype='U')
assert c[0] == uba
assert isinstance(c[:0], table.Column)
assert isinstance(c[0], str)
assert np.all(c[:2] == np.array([uba, 'def']))
assert isinstance(c[:], table.Column)
assert c[:].dtype.char == 'U'
ok = c == [uba, 'def']
assert type(ok) == np.ndarray
assert ok.dtype.char == '?'
assert np.all(ok)
> assert np.all(c != [uba8, b'def'])
astropy/table/tests/test_column.py:777:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <Column dtype='str3' length=2>
bä
def, other = [b'b\xc3\xa4', b'def']
def _compare(self, other):
op = oper # copy enclosed ref to allow swap below
# Special case to work around #6838. Other combinations work OK,
# see tests.test_column.test_unicode_sandwich_compare(). In this
# case just swap self and other.
#
# This is related to an issue in numpy that was addressed in np 1.13.
# However that fix does not make this problem go away, but maybe
# future numpy versions will do so. NUMPY_LT_1_13 to get the
# attention of future maintainers to check (by deleting or versioning
# the if block below). See #6899 discussion.
# 2019-06-21: still needed with numpy 1.16.
if (isinstance(self, MaskedColumn) and self.dtype.kind == 'U'
and isinstance(other, MaskedColumn) and other.dtype.kind == 'S'):
self, other = other, self
op = swapped_oper
if self.dtype.char == 'S':
other = self._encode_str(other)
# Now just let the regular ndarray.__eq__, etc., take over.
> result = getattr(super(Column, self), op)(other)
E FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
astropy/table/column.py:329: FutureWarning
______________________________________________ test_unicode_sandwich_compare[MaskedColumn-MaskedColumn] _______________________________________________
class1 = <class 'astropy.table.column.MaskedColumn'>, class2 = <class 'astropy.table.column.MaskedColumn'>
@pytest.mark.parametrize('class1', [table.MaskedColumn, table.Column])
@pytest.mark.parametrize('class2', [table.MaskedColumn, table.Column, str, list])
def test_unicode_sandwich_compare(class1, class2):
"""Test that comparing a bytestring Column/MaskedColumn with various
str (unicode) object types gives the expected result. Tests #6838.
"""
obj1 = class1([b'a', b'c'])
if class2 is str:
obj2 = 'a'
elif class2 is list:
obj2 = ['a', 'b']
else:
obj2 = class2(['a', 'b'])
assert np.all((obj1 == obj2) == [True, False])
assert np.all((obj2 == obj1) == [True, False])
assert np.all((obj1 != obj2) == [False, True])
assert np.all((obj2 != obj1) == [False, True])
> assert np.all((obj1 > obj2) == [False, True])
E TypeError: '>' not supported between instances of 'MaskedColumn' and 'MaskedColumn'
astropy/table/tests/test_column.py:857: TypeError
_________________________________________________ test_unicode_sandwich_compare[Column-MaskedColumn] __________________________________________________
class1 = <class 'astropy.table.column.MaskedColumn'>, class2 = <class 'astropy.table.column.Column'>
@pytest.mark.parametrize('class1', [table.MaskedColumn, table.Column])
@pytest.mark.parametrize('class2', [table.MaskedColumn, table.Column, str, list])
def test_unicode_sandwich_compare(class1, class2):
"""Test that comparing a bytestring Column/MaskedColumn with various
str (unicode) object types gives the expected result. Tests #6838.
"""
obj1 = class1([b'a', b'c'])
if class2 is str:
obj2 = 'a'
elif class2 is list:
obj2 = ['a', 'b']
else:
obj2 = class2(['a', 'b'])
assert np.all((obj1 == obj2) == [True, False])
assert np.all((obj2 == obj1) == [True, False])
assert np.all((obj1 != obj2) == [False, True])
assert np.all((obj2 != obj1) == [False, True])
> assert np.all((obj1 > obj2) == [False, True])
E TypeError: '>' not supported between instances of 'MaskedColumn' and 'Column'
astropy/table/tests/test_column.py:857: TypeError
____________________________________________________ test_unicode_sandwich_compare[Column-Column] _____________________________________________________
numpy.core._exceptions._UFuncNoLoopError: ufunc 'equal' did not contain a loop with signature matching types (<class 'numpy.dtype[str_]'>, <class 'numpy.dtype[bytes_]'>) -> None
The above exception was the direct cause of the following exception:
class1 = <class 'astropy.table.column.Column'>, class2 = <class 'astropy.table.column.Column'>
@pytest.mark.parametrize('class1', [table.MaskedColumn, table.Column])
@pytest.mark.parametrize('class2', [table.MaskedColumn, table.Column, str, list])
def test_unicode_sandwich_compare(class1, class2):
"""Test that comparing a bytestring Column/MaskedColumn with various
str (unicode) object types gives the expected result. Tests #6838.
"""
obj1 = class1([b'a', b'c'])
if class2 is str:
obj2 = 'a'
elif class2 is list:
obj2 = ['a', 'b']
else:
obj2 = class2(['a', 'b'])
assert np.all((obj1 == obj2) == [True, False])
> assert np.all((obj2 == obj1) == [True, False])
astropy/table/tests/test_column.py:852:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <Column dtype='str1' length=2>
a
b, other = <Column dtype='bytes1' length=2>
a
c
def _compare(self, other):
op = oper # copy enclosed ref to allow swap below
# Special case to work around #6838. Other combinations work OK,
# see tests.test_column.test_unicode_sandwich_compare(). In this
# case just swap self and other.
#
# This is related to an issue in numpy that was addressed in np 1.13.
# However that fix does not make this problem go away, but maybe
# future numpy versions will do so. NUMPY_LT_1_13 to get the
# attention of future maintainers to check (by deleting or versioning
# the if block below). See #6899 discussion.
# 2019-06-21: still needed with numpy 1.16.
if (isinstance(self, MaskedColumn) and self.dtype.kind == 'U'
and isinstance(other, MaskedColumn) and other.dtype.kind == 'S'):
self, other = other, self
op = swapped_oper
if self.dtype.char == 'S':
other = self._encode_str(other)
# Now just let the regular ndarray.__eq__, etc., take over.
> result = getattr(super(Column, self), op)(other)
E FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
astropy/table/column.py:329: FutureWarning
___________________________________________________ test_unicode_sandwich_compare[str-MaskedColumn] ___________________________________________________
class1 = <class 'astropy.table.column.MaskedColumn'>, class2 = <class 'str'>
@pytest.mark.parametrize('class1', [table.MaskedColumn, table.Column])
@pytest.mark.parametrize('class2', [table.MaskedColumn, table.Column, str, list])
def test_unicode_sandwich_compare(class1, class2):
"""Test that comparing a bytestring Column/MaskedColumn with various
str (unicode) object types gives the expected result. Tests #6838.
"""
obj1 = class1([b'a', b'c'])
if class2 is str:
obj2 = 'a'
elif class2 is list:
obj2 = ['a', 'b']
else:
obj2 = class2(['a', 'b'])
assert np.all((obj1 == obj2) == [True, False])
assert np.all((obj2 == obj1) == [True, False])
assert np.all((obj1 != obj2) == [False, True])
assert np.all((obj2 != obj1) == [False, True])
> assert np.all((obj1 > obj2) == [False, True])
E TypeError: '>' not supported between instances of 'MaskedColumn' and 'str'
astropy/table/tests/test_column.py:857: TypeError
__________________________________________________ test_unicode_sandwich_compare[list-MaskedColumn] ___________________________________________________
class1 = <class 'astropy.table.column.MaskedColumn'>, class2 = <class 'list'>
@pytest.mark.parametrize('class1', [table.MaskedColumn, table.Column])
@pytest.mark.parametrize('class2', [table.MaskedColumn, table.Column, str, list])
def test_unicode_sandwich_compare(class1, class2):
"""Test that comparing a bytestring Column/MaskedColumn with various
str (unicode) object types gives the expected result. Tests #6838.
"""
obj1 = class1([b'a', b'c'])
if class2 is str:
obj2 = 'a'
elif class2 is list:
obj2 = ['a', 'b']
else:
obj2 = class2(['a', 'b'])
assert np.all((obj1 == obj2) == [True, False])
assert np.all((obj2 == obj1) == [True, False])
assert np.all((obj1 != obj2) == [False, True])
assert np.all((obj2 != obj1) == [False, True])
> assert np.all((obj1 > obj2) == [False, True])
E TypeError: '>' not supported between instances of 'MaskedColumn' and 'list'
astropy/table/tests/test_column.py:857: TypeError
=============================================================== short test summary info ===============================================================
FAILED astropy/table/tests/test_column.py::test_col_unicode_sandwich_unicode - FutureWarning: elementwise comparison failed; returning scalar instea...
FAILED astropy/table/tests/test_column.py::test_unicode_sandwich_compare[MaskedColumn-MaskedColumn] - TypeError: '>' not supported between instances...
FAILED astropy/table/tests/test_column.py::test_unicode_sandwich_compare[Column-MaskedColumn] - TypeError: '>' not supported between instances of 'M...
FAILED astropy/table/tests/test_column.py::test_unicode_sandwich_compare[Column-Column] - FutureWarning: elementwise comparison failed; returning sc...
FAILED astropy/table/tests/test_column.py::test_unicode_sandwich_compare[str-MaskedColumn] - TypeError: '>' not supported between instances of 'Mask...
FAILED astropy/table/tests/test_column.py::test_unicode_sandwich_compare[list-MaskedColumn] - TypeError: '>' not supported between instances of 'Mas...
=============================================== 6 failed, 3377 passed, 43 skipped, 14 xfailed in 25.62s ===============================================
@pllim - those other errors are actually due to a bug in
Quantity, where the unit of aninitialargument is not taken into account (and where units are no longer stripped in numpy). Working on a fix…I actually don’t know that
-devwas changed too - I think they just reverted the bad commit from 1.23, with the idea that for 1.24 there would be a fix (IIRC, https://github.com/numpy/numpy/pull/21812 would solve at least some of the problems)@pllim - with the corrections to the rc3, i.e., numpy 1.23.x (1.23.0rc3+10.gcc0e08d20), the failures in
io.fits,io.misc, andtableare all gone – all tests pass! So, we can now move to address the problems innumpy-dev.There you go: https://github.com/numpy/numpy/issues/21758
ahh, good point, I forgot that the “nightly” is not in fact a daily build, that at least takes the confusion away of how a partial backport could happen that makes the RC fail but the dev still pass.