Skip to content

Commit 4488f18

Browse files
sinhrksjreback
sinhrks
authored and
jreback
committed
BUG/CLN: datetimelike Index.equals may return True with non-Index
closes #13107 Author: sinhrks <[email protected]> Closes #13986 from sinhrks/dti_equals and squashes the following commits: 580151a [sinhrks] BUG/CLN: move .equals to DatetimeOpsMixin
1 parent e9c5c2d commit 4488f18

18 files changed

+177
-84
lines changed

‎doc/source/whatsnew/v0.19.0.txt

+1
Original file line numberDiff line numberDiff line change
@@ -1461,6 +1461,7 @@ Bug Fixes
14611461
- Bug in operations on ``NaT`` returning ``float`` instead of ``datetime64[ns]`` (:issue:`12941`)
14621462
- Bug in ``Series`` flexible arithmetic methods (like ``.add()``) raises ``ValueError`` when ``axis=None`` (:issue:`13894`)
14631463
- Bug in ``DataFrame.to_csv()`` with ``MultiIndex`` columns in which a stray empty line was added (:issue:`6618`)
1464+
- Bug in ``DatetimeIndex``, ``TimedeltaIndex`` and ``PeriodIndex.equals()`` may return ``True`` when input isn't ``Index`` but contains the same values (:issue:`13107`)
14641465

14651466

14661467
- Bug in ``Index`` raises ``KeyError`` displaying incorrect column when column is not in the df and columns contains duplicate values (:issue:`13822`)

‎pandas/indexes/base.py

+9-2
Original file line numberDiff line numberDiff line change
@@ -1605,8 +1605,15 @@ def equals(self, other):
16051605
if not isinstance(other, Index):
16061606
return False
16071607

1608-
return array_equivalent(_values_from_object(self),
1609-
_values_from_object(other))
1608+
if is_object_dtype(self) and not is_object_dtype(other):
1609+
# if other is not object, use other's logic for coercion
1610+
return other.equals(self)
1611+
1612+
try:
1613+
return array_equivalent(_values_from_object(self),
1614+
_values_from_object(other))
1615+
except:
1616+
return False
16101617

16111618
def identical(self, other):
16121619
"""Similar to equals, but check that other comparable attributes are

‎pandas/indexes/category.py

+3
Original file line numberDiff line numberDiff line change
@@ -196,6 +196,9 @@ def equals(self, other):
196196
if self.is_(other):
197197
return True
198198

199+
if not isinstance(other, Index):
200+
return False
201+
199202
try:
200203
other = self._is_dtype_compat(other)
201204
return array_equivalent(self._data, other)

‎pandas/indexes/multi.py

+4
Original file line numberDiff line numberDiff line change
@@ -1436,6 +1436,7 @@ def reindex(self, target, method=None, level=None, limit=None,
14361436
return_indexers=True,
14371437
keep_order=False)
14381438
else:
1439+
target = _ensure_index(target)
14391440
if self.equals(target):
14401441
indexer = None
14411442
else:
@@ -1984,6 +1985,9 @@ def equals(self, other):
19841985
if self.is_(other):
19851986
return True
19861987

1988+
if not isinstance(other, Index):
1989+
return False
1990+
19871991
if not isinstance(other, MultiIndex):
19881992
return array_equivalent(self._values,
19891993
_values_from_object(_ensure_index(other)))

‎pandas/indexes/numeric.py

+4-11
Original file line numberDiff line numberDiff line change
@@ -7,7 +7,7 @@
77
from pandas.types.common import (is_dtype_equal, pandas_dtype,
88
is_float_dtype, is_object_dtype,
99
is_integer_dtype, is_scalar)
10-
from pandas.types.missing import array_equivalent, isnull
10+
from pandas.types.missing import isnull
1111
from pandas.core.common import _values_from_object
1212

1313
from pandas import compat
@@ -160,16 +160,6 @@ def _convert_scalar_indexer(self, key, kind=None):
160160
return (super(Int64Index, self)
161161
._convert_scalar_indexer(key, kind=kind))
162162

163-
def equals(self, other):
164-
"""
165-
Determines if two Index objects contain the same elements.
166-
"""
167-
if self.is_(other):
168-
return True
169-
170-
return array_equivalent(_values_from_object(self),
171-
_values_from_object(other))
172-
173163
def _wrap_joined_index(self, joined, other):
174164
name = self.name if self.name == other.name else None
175165
return Int64Index(joined, name=name)
@@ -306,6 +296,9 @@ def equals(self, other):
306296
if self is other:
307297
return True
308298

299+
if not isinstance(other, Index):
300+
return False
301+
309302
# need to compare nans locations and make sure that they are the same
310303
# since nans don't compare equal this is a bit tricky
311304
try:

‎pandas/tests/indexes/common.py

+14
Original file line numberDiff line numberDiff line change
@@ -650,6 +650,20 @@ def test_delete_base(self):
650650
# either depending on numpy version
651651
result = idx.delete(len(idx))
652652

653+
def test_equals(self):
654+
655+
for name, idx in compat.iteritems(self.indices):
656+
self.assertTrue(idx.equals(idx))
657+
self.assertTrue(idx.equals(idx.copy()))
658+
self.assertTrue(idx.equals(idx.astype(object)))
659+
660+
self.assertFalse(idx.equals(list(idx)))
661+
self.assertFalse(idx.equals(np.array(idx)))
662+
663+
if idx.nlevels == 1:
664+
# do not test MultiIndex
665+
self.assertFalse(idx.equals(pd.Series(idx)))
666+
653667
def test_equals_op(self):
654668
# GH9947, GH10637
655669
index_a = self.create_index()

‎pandas/tests/indexes/test_base.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -400,7 +400,7 @@ def test_astype(self):
400400
casted = self.intIndex.astype('i8')
401401
self.assertEqual(casted.name, 'foobar')
402402

403-
def test_equals(self):
403+
def test_equals_object(self):
404404
# same
405405
self.assertTrue(Index(['a', 'b', 'c']).equals(Index(['a', 'b', 'c'])))
406406

‎pandas/tests/indexes/test_category.py

+24-13
Original file line numberDiff line numberDiff line change
@@ -522,7 +522,7 @@ def test_ensure_copied_data(self):
522522
result = CategoricalIndex(index.values, copy=False)
523523
self.assertIs(_base(index.values), _base(result.values))
524524

525-
def test_equals(self):
525+
def test_equals_categorical(self):
526526

527527
ci1 = CategoricalIndex(['a', 'b'], categories=['a', 'b'], ordered=True)
528528
ci2 = CategoricalIndex(['a', 'b'], categories=['a', 'b', 'c'],
@@ -556,19 +556,30 @@ def test_equals(self):
556556

557557
# tests
558558
# make sure that we are testing for category inclusion properly
559-
self.assertTrue(CategoricalIndex(
560-
list('aabca'), categories=['c', 'a', 'b']).equals(list('aabca')))
559+
ci = CategoricalIndex(list('aabca'), categories=['c', 'a', 'b'])
560+
self.assertFalse(ci.equals(list('aabca')))
561+
self.assertFalse(ci.equals(CategoricalIndex(list('aabca'))))
562+
self.assertTrue(ci.equals(ci.copy()))
563+
564+
with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
565+
ci = CategoricalIndex(list('aabca'),
566+
categories=['c', 'a', 'b', np.nan])
567+
self.assertFalse(ci.equals(list('aabca')))
568+
self.assertFalse(ci.equals(CategoricalIndex(list('aabca'))))
561569
with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
562-
self.assertTrue(CategoricalIndex(
563-
list('aabca'), categories=['c', 'a', 'b', np.nan]).equals(list(
564-
'aabca')))
565-
566-
self.assertFalse(CategoricalIndex(
567-
list('aabca') + [np.nan], categories=['c', 'a', 'b']).equals(list(
568-
'aabca')))
569-
self.assertTrue(CategoricalIndex(
570-
list('aabca') + [np.nan], categories=['c', 'a', 'b']).equals(list(
571-
'aabca') + [np.nan]))
570+
self.assertTrue(ci.equals(ci.copy()))
571+
572+
ci = CategoricalIndex(list('aabca') + [np.nan],
573+
categories=['c', 'a', 'b'])
574+
self.assertFalse(ci.equals(list('aabca')))
575+
self.assertFalse(ci.equals(CategoricalIndex(list('aabca'))))
576+
self.assertTrue(ci.equals(ci.copy()))
577+
578+
ci = CategoricalIndex(list('aabca') + [np.nan],
579+
categories=['c', 'a', 'b'])
580+
self.assertFalse(ci.equals(list('aabca') + [np.nan]))
581+
self.assertFalse(ci.equals(CategoricalIndex(list('aabca') + [np.nan])))
582+
self.assertTrue(ci.equals(ci.copy()))
572583

573584
def test_string_categorical_index_repr(self):
574585
# short

‎pandas/tests/indexes/test_multi.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -1266,7 +1266,7 @@ def test_to_hierarchical(self):
12661266
def test_bounds(self):
12671267
self.index._bounds
12681268

1269-
def test_equals(self):
1269+
def test_equals_multi(self):
12701270
self.assertTrue(self.index.equals(self.index))
12711271
self.assertTrue(self.index.equal_levels(self.index))
12721272

‎pandas/tests/indexes/test_numeric.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -265,7 +265,7 @@ def test_astype(self):
265265
i = Float64Index([0, 1.1, np.NAN])
266266
self.assertRaises(ValueError, lambda: i.astype(dtype))
267267

268-
def test_equals(self):
268+
def test_equals_numeric(self):
269269

270270
i = Float64Index([1.0, 2.0])
271271
self.assertTrue(i.equals(i))

‎pandas/tests/indexes/test_range.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -337,7 +337,7 @@ def test_is_monotonic(self):
337337
self.assertTrue(index.is_monotonic_increasing)
338338
self.assertTrue(index.is_monotonic_decreasing)
339339

340-
def test_equals(self):
340+
def test_equals_range(self):
341341
equiv_pairs = [(RangeIndex(0, 9, 2), RangeIndex(0, 10, 2)),
342342
(RangeIndex(0), RangeIndex(1, -1, 3)),
343343
(RangeIndex(1, 2, 3), RangeIndex(1, 3, 4)),

‎pandas/tseries/base.py

+29-1
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
import numpy as np
1212
from pandas.types.common import (is_integer, is_float,
1313
is_bool_dtype, _ensure_int64,
14-
is_scalar,
14+
is_scalar, is_dtype_equal,
1515
is_list_like)
1616
from pandas.types.generic import (ABCIndex, ABCSeries,
1717
ABCPeriodIndex, ABCIndexClass)
@@ -109,6 +109,34 @@ def ceil(self, freq):
109109
class DatetimeIndexOpsMixin(object):
110110
""" common ops mixin to support a unified inteface datetimelike Index """
111111

112+
def equals(self, other):
113+
"""
114+
Determines if two Index objects contain the same elements.
115+
"""
116+
if self.is_(other):
117+
return True
118+
119+
if not isinstance(other, ABCIndexClass):
120+
return False
121+
elif not isinstance(other, type(self)):
122+
try:
123+
other = type(self)(other)
124+
except:
125+
return False
126+
127+
if not is_dtype_equal(self.dtype, other.dtype):
128+
# have different timezone
129+
return False
130+
131+
# ToDo: Remove this when PeriodDtype is added
132+
elif isinstance(self, ABCPeriodIndex):
133+
if not isinstance(other, ABCPeriodIndex):
134+
return False
135+
if self.freq != other.freq:
136+
return False
137+
138+
return np.array_equal(self.asi8, other.asi8)
139+
112140
def __iter__(self):
113141
return (self._box_func(v) for v in self.asi8)
114142

‎pandas/tseries/index.py

-20
Original file line numberDiff line numberDiff line change
@@ -1625,26 +1625,6 @@ def is_normalized(self):
16251625
def _resolution(self):
16261626
return period.resolution(self.asi8, self.tz)
16271627

1628-
def equals(self, other):
1629-
"""
1630-
Determines if two Index objects contain the same elements.
1631-
"""
1632-
if self.is_(other):
1633-
return True
1634-
1635-
if (not hasattr(other, 'inferred_type') or
1636-
other.inferred_type != 'datetime64'):
1637-
if self.offset is not None:
1638-
return False
1639-
try:
1640-
other = DatetimeIndex(other)
1641-
except:
1642-
return False
1643-
1644-
if self._has_same_tz(other):
1645-
return np.array_equal(self.asi8, other.asi8)
1646-
return False
1647-
16481628
def insert(self, loc, item):
16491629
"""
16501630
Make new Index inserting new item at location

‎pandas/tseries/period.py

-15
Original file line numberDiff line numberDiff line change
@@ -596,21 +596,6 @@ def _mpl_repr(self):
596596
# how to represent ourselves to matplotlib
597597
return self.asobject.values
598598

599-
def equals(self, other):
600-
"""
601-
Determines if two Index objects contain the same elements.
602-
"""
603-
if self.is_(other):
604-
return True
605-
606-
if not isinstance(other, PeriodIndex):
607-
try:
608-
other = PeriodIndex(other)
609-
except:
610-
return False
611-
612-
return np.array_equal(self.asi8, other.asi8)
613-
614599
def to_timestamp(self, freq=None, how='start'):
615600
"""
616601
Cast to DatetimeIndex

‎pandas/tseries/tdi.py

-16
Original file line numberDiff line numberDiff line change
@@ -806,22 +806,6 @@ def dtype(self):
806806
def is_all_dates(self):
807807
return True
808808

809-
def equals(self, other):
810-
"""
811-
Determines if two Index objects contain the same elements.
812-
"""
813-
if self.is_(other):
814-
return True
815-
816-
if (not hasattr(other, 'inferred_type') or
817-
other.inferred_type != 'timedelta64'):
818-
try:
819-
other = TimedeltaIndex(other)
820-
except:
821-
return False
822-
823-
return np.array_equal(self.asi8, other.asi8)
824-
825809
def insert(self, loc, item):
826810
"""
827811
Make new Index inserting new item at location

0 commit comments

Comments
 (0)