Skip to content

BUG: Concatenation of DFs with all NaT columns and TZ-aware ones breaks #12396 #12403

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

Closed
wants to merge 3 commits into from
Closed

BUG: Concatenation of DFs with all NaT columns and TZ-aware ones breaks #12396 #12403

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
44a0fcf
handle tz-aware all NaT concatenation incl. mixed cases (GH12396)
multiloc May 8, 2016
f95bb83
improve tests for all-NaT concatenation (GH12396)
May 8, 2016
10406df
clean up tests for GH12396
May 8, 2016
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
6 changes: 6 additions & 0 deletions pandas/core/internals.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4911,6 +4911,12 @@ def get_reindexed_values(self, empty_dtype, upcasted_na):
pass
elif getattr(self.block, 'is_sparse', False):
pass
elif com.is_extension_type(empty_dtype) and \
com.is_datetimetz(empty_dtype):
num_elements = np.prod(self.shape)
# handle timezone-aware all NaT cases
Copy link
Contributor

Choose a reason for hiding this comment

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

no, do this under the if getattr(self.block, 'is_datetimetz', False) or com.is_datetimetz(empty_dtype)

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Like so? This passes the tests but not sure how good the coverage is for the non-NaT tz-aware cases.

                if getattr(self.block, 'is_datetimetz', False) \
                    or com.is_datetimetz(empty_dtype):
                    num_elements = np.prod(self.shape)
                    # handle timezone-aware all NaT cases
                    return DatetimeIndex([fill_value] * num_elements,
                                         dtype=empty_dtype)
                elif getattr(self.block, 'is_categorical', False):
                    pass
                elif getattr(self.block, 'is_sparse', False):
                    pass

Copy link
Contributor

Choose a reason for hiding this comment

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

yes more like this. yes coverage of extension types is prob only so now

return DatetimeIndex([fill_value] * num_elements,
dtype=empty_dtype)
else:
missing_arr = np.empty(self.shape, dtype=empty_dtype)
missing_arr.fill(fill_value)
Expand Down
115 changes: 115 additions & 0 deletions pandas/tools/tests/test_merge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@
from pandas.compat import range, lrange, lzip, StringIO
from pandas import compat
from pandas.tseries.index import DatetimeIndex
from pandas.types.dtypes import DatetimeTZDtype
from pandas.tools.merge import merge, concat, ordered_merge, MergeError
from pandas import Categorical, Timestamp
from pandas.util.testing import (assert_frame_equal, assert_series_equal,
Expand Down Expand Up @@ -2522,6 +2523,120 @@ def test_concat_multiindex_with_tz(self):
result = concat([df, df])
tm.assert_frame_equal(result, expected)

def test_concat_NaT_dataframes_all_NaT_axis_0(self):
# GH 12396
expect = pd.DataFrame([pd.NaT, pd.NaT, pd.NaT], index=[0, 1, 0])

# non-timezone aware
first = pd.DataFrame([[pd.NaT], [pd.NaT]])
second = pd.DataFrame([[pd.NaT]])

result = pd.concat([first, second], axis=0)
assert_frame_equal(result, expect)

# one side timezone-aware
dtype = DatetimeTZDtype('ns', tz='UTC')
first = pd.DataFrame([[pd.NaT], [pd.NaT]], dtype=dtype)

result = pd.concat([first, second], axis=0)
# upcasts for mixed case
assert_frame_equal(result, expect, check_dtype=False)
Copy link
Contributor

Choose a reason for hiding this comment

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

make a new expect here that matches exactly what you want rather, so assert_frame_equal(result, expected) works & checks dtypes

self.assertEqual(result.dtypes[0], dtype)

# both sides timezone-aware
second = pd.DataFrame([[pd.NaT]], dtype=dtype)

result = pd.concat([first, second], axis=0)
# upcasts to tz-aware
assert_frame_equal(result, expect, check_dtype=False)
self.assertEqual(result.dtypes[0], dtype)
Copy link
Contributor

Choose a reason for hiding this comment

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

same here


def test_concat_NaT_dataframes_all_NaT_axis_1(self):
# GH 12396
expect = pd.DataFrame([[pd.NaT, pd.NaT], [pd.NaT, pd.NaT]],
columns=[0, 0])

# non-timezone aware
first = pd.DataFrame([[pd.NaT], [pd.NaT]])
second = pd.DataFrame([[pd.NaT]])

result = pd.concat([first, second], axis=1)
assert_frame_equal(result, expect)

# one side timezone-aware
dtype = DatetimeTZDtype('ns', tz='UTC')
first = pd.DataFrame([[pd.NaT], [pd.NaT]], dtype=dtype)

# upcasts result to tz-aware
assert_frame_equal(result, expect, check_dtype=False)
Copy link
Contributor

Choose a reason for hiding this comment

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

same here, make an exactly frame of what you are comparing.

result = pd.concat([first, second], axis=1)
self.assertEqual(result.dtypes.iloc[0], dtype)
self.assertEqual(result.dtypes.iloc[0], first.dtypes[0])
self.assertEqual(result.dtypes.iloc[1], second.dtypes[0])

# both sides timezone-aware
second = pd.DataFrame([[pd.NaT]], dtype=dtype)

result = pd.concat([first, second], axis=1)
assert_frame_equal(result, expect, check_dtype=False)
# upcasts to tz-aware
self.assertEqual(result.dtypes.iloc[0], dtype)
self.assertEqual(result.dtypes.iloc[0], first.dtypes[0])
Copy link
Contributor

Choose a reason for hiding this comment

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

same

self.assertEqual(result.dtypes.iloc[1], second.dtypes[0])

def test_concat_NaT_dataframes_mixed_timestamps_and_NaT(self):
# GH 12396

# non-timezone aware
first = pd.DataFrame([[pd.NaT], [pd.NaT]])
second = pd.DataFrame([[pd.Timestamp('2015/01/01')],
[pd.Timestamp('2016/01/01')]])

expect = pd.DataFrame([pd.NaT, pd.NaT, second.iloc[0, 0],
second.iloc[1, 0]], index=[0, 1, 0, 1])

result = pd.concat([first, second], axis=0)
assert_frame_equal(result, expect)
self.assertEqual(result.dtypes.iloc[0], first.dtypes[0])

# one side timezone-aware
dtype = DatetimeTZDtype('ns', tz='UTC')
second = second.apply(lambda x: x.astype(dtype))

result = pd.concat([first, second], axis=0)
assert_frame_equal(result, expect, check_dtype=False)
# upcasts
self.assertEqual(result.dtypes.iloc[0], dtype)
self.assertEqual(result.dtypes.iloc[0], second.dtypes[0])

def test_concat_NaT_series_dataframe_all_NaT(self):
# GH 12396

# non-timezone aware
first = pd.Series([pd.NaT, pd.NaT])
second = pd.DataFrame([[pd.Timestamp('2015/01/01')],
[pd.Timestamp('2016/01/01')]])

expect = pd.DataFrame([pd.NaT, pd.NaT, second.iloc[0, 0],
second.iloc[1, 0]], index=[0, 1, 0, 1])

result = pd.concat([first, second])
assert_frame_equal(result, expect)

# one side timezone-aware
dtype = DatetimeTZDtype('ns', tz='UTC')
second = second.apply(lambda x: x.astype(dtype))

result = pd.concat([first, second])

expect = expect.apply(lambda x: x.astype(dtype))
assert_frame_equal(result, expect, check_dtype=True)

# both sides timezone-aware
first = first.astype(dtype)
result = pd.concat([first, second])
assert_frame_equal(result, expect, check_dtype=True)

def test_concat_keys_and_levels(self):
df = DataFrame(np.random.randn(1, 3))
df2 = DataFrame(np.random.randn(1, 4))
Expand Down