Skip to content

TST: Added test case for DataFrame.at #33577

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 5 commits into from
Apr 21, 2020
Merged

TST: Added test case for DataFrame.at #33577

Changes from 3 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
b2e4013
Added test case from issue
EdAbati Apr 15, 2020
18d727f
Changed assert to equal utc
EdAbati Apr 15, 2020
fdbfa82
fixed imports with isort
EdAbati Apr 16, 2020
ac97198
switched to assert_frame_equal
EdAbati Apr 17, 2020
13fdfbe
fixed dytpe in expected dataframe
EdAbati Apr 17, 2020
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
10 changes: 10 additions & 0 deletions pandas/tests/frame/indexing/test_at.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
from datetime import datetime, timezone

import pandas as pd


def test_at_timezone():
# https://github.com/pandas-dev/pandas/issues/33544
df = pd.DataFrame({"foo": [datetime(2000, 1, 1)]})
df.at[0, "foo"] = datetime(2000, 1, 2, tzinfo=timezone.utc)
assert df.at[0, "foo"].tzinfo == timezone.utc
Copy link
Member

Choose a reason for hiding this comment

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

Could you instead create an expected Dataframe and use tm.assert_frame_equal?

    result = pd.DataFrame({"foo": [datetime(2000, 1, 1)]})
    result.at[0, "foo"] = datetime(2000, 1, 2, tzinfo=timezone.utc)
    expected = pd.DataFrame(....)
    tm.assert_frame_equal(result, expected)

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Hi @mroeschke ,

I have implemented as you suggested, but actually (as you can see in the checks below) this makes the test fail and raises an interesting question.

When creating the dataframe with:

result = pd.DataFrame({"foo": [datetime(2000, 1, 1)]})

It transforms the value result.at[0, "foo"] into a Timestamp object:Out[4]: Timestamp('2000-01-01 00:00:00')and running result.info() will show that the column Dtype is datetime64[ns].

If now I apply the following

result.at[0, "foo"] = datetime(2000, 1, 2, tzinfo=timezone.utc)

the result.at[0, "foo"] is transformed in to a datetime object
Out[6]: datetime.datetime(2000, 1, 2, 0, 0, tzinfo=datetime.timezone.utc) and also the column Dtype is changed to object.

I am not sure if this is the expected behaviour or it is a bug.

If this is the expected behaviour I should change the test, otherwise .at needs to be fixed.

Copy link
Member

Choose a reason for hiding this comment

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

I think this is the expected behavior as opposed to the loc variant that would coerce to datetime type.

You can therefore change the expected construction to

expected = pd.DataFrame(np.array([datetime(2000, 1, 2, tzinfo=timezone.utc)], dtype=object, columns['foo'])

Copy link
Member

Choose a reason for hiding this comment

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

I think this is the expected behavior

this seems inconsisent with other types that cannot hold the scalar type.

>>> import numpy as np
>>>
>>> import pandas as pd
>>>
>>> pd.__version__
'1.1.0.dev0+1290.g7e99bb433'
>>>
>>> df = pd.DataFrame({"a": [1, 2, 3]})
>>> df
   a
0  1
1  2
2  3
>>>
>>> df.at[0, "a"] = np.nan
Traceback (most recent call last):
...
ValueError: Cannot assign nan to integer series
>>>