BUG: df.where() inconsistently casts columns to integers

[nicolaslegrand91]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Code Sample, a copy-pastable example

import pandas as pd
import numpy as np

# Float column casted to int
d1 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, np.nan]})
print('\nd1:')
print(d1)
print('\nd1.dtypes:')
print(d1.dtypes)
print('\nd1.where(pd.notnull(d1), None):')
print(d1.where(pd.notnull(d1), None))

print()

# Float column not casted to int
d2 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, 4]})
print('\nd2:')
print(d2)
print('\nd2.dtypes:')
print(d2.dtypes)
print('\nd2.where(pd.notnull(d2), None):')
print(d2.where(pd.notnull(d2), None))

Problem description

In both previous examples, the column a is identical (a float column containing 1.0 and 2.0), but the outcome of df.where(pd.notnull(df), None) differs.

Expected Output

For a given column, I would expect the outcome of df.where(pd.notnull(df), None) to be consistent, regardless of the other columns of the dataframe.

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 2cb9652
python : 3.9.5.final.0
python-bits : 64
OS : Darwin
OS-release : 19.6.0
Version : Darwin Kernel Version 19.6.0: Tue Jan 12 22:13:05 PST 2021; root:xnu-6153.141.16~1/RELEASE_X86_64
machine : x86_64
processor : i386
byteorder : little
LC_ALL : en_US.UTF-8
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.2.4
numpy : 1.20.3
pytz : 2021.1
dateutil : 2.8.1
pip : 21.1.2
setuptools : 49.6.0
Cython : 0.29.23
pytest : 6.2.3
hypothesis : None
sphinx : 3.5.4
blosc : None
feather : None
xlsxwriter : 1.4.3
lxml.etree : 4.6.3
html5lib : None
pymysql : 1.0.2
psycopg2 : 2.8.6 (dt dec pq3 ext lo64)
jinja2 : 2.11.3
IPython : 7.24.1
pandas_datareader: None
bs4 : 4.9.3
bottleneck : 1.3.2
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.4.2
numexpr : None
odfpy : None
openpyxl : 3.0.7
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.6.3
sqlalchemy : 1.4.17
tables : None
tabulate : None
xarray : None
xlrd : 2.0.1
xlwt : None
numba : None

[MemphisMeng]

It wasn't cast to integers at all, it was just displayed without .0. But it's still a good issue to work on ;)

[nicolaslegrand91]

@MemphisMeng Thanks for your reply. Actually it seems it was really casted to integers:

import pandas as pd
import numpy as np

# Float column casted to int
d1 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, np.nan]})
print('\nd1:')
print(d1)
print('\nd1.dtypes:')
print(d1.dtypes)
print('\nd1.where(pd.notnull(d1), None):')
print(d1.where(pd.notnull(d1), None))
print('\nd1.where(pd.notnull(d1), None).dtypes:')
print(d1.where(pd.notnull(d1), None).dtypes)

print()

# Float column not casted to int
d2 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, 4]})
print('\nd2:')
print(d2)
print('\nd2.dtypes:')
print(d2.dtypes)
print('\nd2.where(pd.notnull(d2), None):')
print(d2.where(pd.notnull(d2), None))
print('\nd2.where(pd.notnull(d2), None).dtypes:')
print(d2.where(pd.notnull(d2), None).dtypes)

[MemphisMeng]

Alright, I mean when I tried inplace=True, and type the df out afterward, it showed as expected.

[mroeschke]

This looks consistent on master now. Could use a test

In [18]: import pandas as pd
    ...: import numpy as np
    ...:
    ...: # Float column casted to int
    ...: d1 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, np.nan]})
    ...: print('\nd1:')
    ...: print(d1)
    ...: print('\nd1.dtypes:')
    ...: print(d1.dtypes)
    ...: print('\nd1.where(pd.notnull(d1), None):')
    ...: print(d1.where(pd.notnull(d1), None))
    ...:
    ...: print()
    ...:
    ...: # Float column not casted to int
    ...: d2 = pd.DataFrame({"a": [1.0, 2.0], 'b': [3, 4]})
    ...: print('\nd2:')
    ...: print(d2)
    ...: print('\nd2.dtypes:')
    ...: print(d2.dtypes)
    ...: print('\nd2.where(pd.notnull(d2), None):')
    ...: print(d2.where(pd.notnull(d2), None))

d1:
     a    b
0  1.0  3.0
1  2.0  NaN

d1.dtypes:
a    float64
b    float64
dtype: object

d1.where(pd.notnull(d1), None):
     a    b
0  1.0  3.0
1  2.0  NaN


d2:
     a  b
0  1.0  3
1  2.0  4

d2.dtypes:
a    float64
b      int64
dtype: object

d2.where(pd.notnull(d2), None):
     a  b
0  1.0  3
1  2.0  4

[maximilianaccardo]

take

[san-coding]

take