Skip to content

BUG: DataFrame.replace with out of bound datetime causing RecursionError #22108

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Aug 1, 2018
Merged

BUG: DataFrame.replace with out of bound datetime causing RecursionError #22108

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
0213b82
convert=False during exception handling in Block.replace
minggli Jul 28, 2018
42fc85e
try convert datetime
minggli Jul 28, 2018
02e3f2f
refactor and add test case
minggli Jul 28, 2018
fcfd675
remove False convert
minggli Jul 28, 2018
8bdcd62
add dt64 with tzinfo
minggli Jul 29, 2018
e3aba78
restricter exception handling and clearer comments
minggli Jul 29, 2018
cb2a3c6
add whatsnew
minggli Jul 29, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion pandas/core/dtypes/cast.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -838,7 +838,12 @@ def soft_convert_objects(values, datetime=True, numeric=True, timedelta=True,

# Soft conversions
if datetime:
values = lib.maybe_convert_objects(values, convert_datetime=datetime)
# GH 20380
try:
values = lib.maybe_convert_objects(values,
convert_datetime=datetime)
except ValueError:
pass
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you add a comment on when the valueerror is hit

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

commented, used OutOfBoundsDatetime to be more clear on the error


if timedelta and is_object_dtype(values.dtype):
# Object check to ensure only run if previous did not convert
Expand Down
10 changes: 6 additions & 4 deletions pandas/core/internals/blocks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -802,12 +802,14 @@ def replace(self, to_replace, value, inplace=False, filter=None,
copy=not inplace) for b in blocks]
return blocks
except (TypeError, ValueError):

# try again with a compatible block
block = self.astype(object)
return block.replace(
to_replace=original_to_replace, value=value, inplace=inplace,
filter=filter, regex=regex, convert=convert)
return block.replace(to_replace=original_to_replace,
value=value,
inplace=inplace,
filter=filter,
regex=regex,
convert=convert)

def _replace_single(self, *args, **kwargs):
""" no-op on a non-ObjectBlock """
Expand Down
57 changes: 23 additions & 34 deletions pandas/tests/frame/test_replace.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -757,40 +757,29 @@ def test_replace_for_new_dtypes(self):
result = tsframe.fillna(method='bfill')
assert_frame_equal(result, tsframe.fillna(method='bfill'))

def test_replace_dtypes(self):
# int
df = DataFrame({'ints': [1, 2, 3]})
result = df.replace(1, 0)
expected = DataFrame({'ints': [0, 2, 3]})
assert_frame_equal(result, expected)

df = DataFrame({'ints': [1, 2, 3]}, dtype=np.int32)
result = df.replace(1, 0)
expected = DataFrame({'ints': [0, 2, 3]}, dtype=np.int32)
assert_frame_equal(result, expected)

df = DataFrame({'ints': [1, 2, 3]}, dtype=np.int16)
result = df.replace(1, 0)
expected = DataFrame({'ints': [0, 2, 3]}, dtype=np.int16)
assert_frame_equal(result, expected)

# bools
df = DataFrame({'bools': [True, False, True]})
result = df.replace(False, True)
assert result.values.all()

# complex blocks
df = DataFrame({'complex': [1j, 2j, 3j]})
result = df.replace(1j, 0j)
expected = DataFrame({'complex': [0j, 2j, 3j]})
assert_frame_equal(result, expected)

# datetime blocks
prev = datetime.today()
now = datetime.today()
df = DataFrame({'datetime64': Index([prev, now, prev])})
result = df.replace(prev, now)
expected = DataFrame({'datetime64': Index([now] * 3)})
@pytest.mark.parametrize('frame, to_replace, value, expected', [
(DataFrame({'ints': [1, 2, 3]}), 1, 0,
DataFrame({'ints': [0, 2, 3]})),
(DataFrame({'ints': [1, 2, 3]}, dtype=np.int32), 1, 0,
DataFrame({'ints': [0, 2, 3]}, dtype=np.int32)),
(DataFrame({'ints': [1, 2, 3]}, dtype=np.int16), 1, 0,
DataFrame({'ints': [0, 2, 3]}, dtype=np.int16)),
(DataFrame({'bools': [True, False, True]}), False, True,
DataFrame({'bools': [True, True, True]})),
(DataFrame({'complex': [1j, 2j, 3j]}), 1j, 0,
DataFrame({'complex': [0j, 2j, 3j]})),
(DataFrame({'datetime64': Index([datetime(2018, 5, 28),
datetime(2018, 7, 28),
datetime(2018, 5, 28)])}),
datetime(2018, 5, 28), datetime(2018, 7, 28),
DataFrame({'datetime64': Index([datetime(2018, 7, 28)] * 3)})),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you add a datetime64 w/tz here as well (if its an easy add)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sure! 👍

# GH 20380
(DataFrame({'dt': [datetime(3017, 12, 20)], 'str': ['foo']}),
'foo', 'bar',
DataFrame({'dt': [datetime(3017, 12, 20)], 'str': ['bar']}))
])
def test_replace_dtypes(self, frame, to_replace, value, expected):
result = getattr(frame, 'replace')(to_replace, value)
assert_frame_equal(result, expected)

def test_replace_input_formats_listlike(self):
Expand Down