DataFrame.at indexing with date string promotes column type, while indexing with datetime casts value

[Jokab]

Code Sample, a copy-pastable example if possible

index = pd.date_range("2019-01-01", "2019-01-02", freq="D")
df = pd.DataFrame(index=index, columns=["A"]).fillna(1)

>>> df.at[pd.to_datetime('2019-01-01'), 'A'] = 0.98
>>> df
            A
2019-01-01  0  # cast to column dtype
2019-01-02  1
>>> df['A'].dtype
dtype('int64')

>>> df.at['2019-01-01', 'A'] = 0.98
>>> df
               A
2019-01-01  0.98  # column cast to value dtype
2019-01-02  1.00
>>> df['A'].dtype
dtype('float64')

Problem description

This issue arises when dealing with a Dataframe with a datetime index. When using DataFrame.at[] and indexing using a date string, the column dtype is promoted to the type of the provided value. However when using DataFrame.at[] and indexing using a datetime, the provided value is instead cast to the column's dtype. In the example, 0.98 is in this case indeed cast to 0.

Expected Output

That indexing DataFrame.at[] with either a date string or a datetime works the same.

Output of pd.show_versions()

[paste the output of pd.show_versions() here below this line]

pd.show_versions()
INSTALLED VERSIONS

---

commit : None
python : 3.6.8.final.0
python-bits : 64
OS : Darwin
OS-release : 18.2.0
machine : x86_64
processor : i386
byteorder : little
LC_ALL : en_US.UTF-8
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 0.25.3
numpy : 1.16.1
pytz : 2019.3
dateutil : 2.8.0
pip : 19.3.1
setuptools : 41.6.0
Cython : None
pytest : 4.3.0
hypothesis : 4.7.11
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : 2.7.7 (dt dec pq3 ext lo64)
jinja2 : 2.10.3
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
s3fs : None
scipy : 1.3.1
sqlalchemy : None
tables : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None

[martinwrudolf]

Hi, can I take on this issue?

[jbrockmendel]

This works on master, best guess is closed by #45121. Will add a test for it.

[1nF0rmed]

@jbrockmendel were you able to add a test for this or is this issue still open?

[jbrockmendel]

not that i can recall

[AryanAgarwal2000]

Which file would be ideal to write the test in? Is pandas/tests the correct folder?

[nealxm]

hello @AryanAgarwal2000 are you still working on this? if so, i think you could add your code to pandas/tests/indexing/test_at.py. if not, i would like to work on this issue.