BUG: Change in index of empty dataframes in mode operation

[galipremsagar]

• 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.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

Previously pandas used to retain the index in an empty dataframe:

>>> df = pd.DataFrame({'a':['a', 'b', 'c']}, index=['a', 'b', 'c'])
>>> df.mode()
   a
0  a
1  b
2  c
>>> df.mode(numeric_only=True)
Empty DataFrame
Columns: []
Index: [a, b, c]
>>> pd.__version__
'1.2.5'

Currently the index is not being retained if the dataframe is empty:

>>> import pandas as pd
>>> df = pd.DataFrame({'a':['a', 'b', 'c']}, index=['a', 'b', 'c'])
>>> df.mode()
   a
0  a
1  b
2  c
>>> df.mode(numeric_only=True)
Empty DataFrame
Columns: []
Index: []
>>> pd.__version__

Is this a conscious change, if yes can someone point me to the docs where I can learn more about this change?

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 5f648bf
python : 3.8.10.final.0
python-bits : 64
OS : Linux
OS-release : 4.15.0-76-generic
Version : #86-Ubuntu SMP Fri Jan 17 17:24:28 UTC 2020
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.3.2
numpy : 1.21.2
pytz : 2021.1
dateutil : 2.8.2
pip : 21.2.4
setuptools : 57.4.0
Cython : 0.29.24
pytest : 6.2.5
hypothesis : 6.17.4
sphinx : 4.1.2
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.0.1
IPython : 7.27.0
pandas_datareader: None
bs4 : 4.9.3
bottleneck : None
fsspec : 2021.08.1
fastparquet : None
gcsfs : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 5.0.0
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : 0.53.1

[simonjayhawkins]

Thanks @galipremsagar for the report.

The new behavior looks potentially correct to me. (It maybe should raise instead @jbrockmendel?)

The number of rows in the result of a mode operation (with the default axis) corresponds to the highest number of modes in the results over all the columns.

For the first example, there are 3 modes in the only column present, hence 3 rows in the result.

In the second example there are no columns to evaluate (since numeric_only=True) and hence no modes so I would expect the result to have no rows.

Is this a conscious change, if yes can someone point me to the docs where I can learn more about this change?

I do a bisect shortly to get more context.

[simonjayhawkins]

Is this a conscious change, if yes can someone point me to the docs where I can learn more about this change?

I do a bisect shortly to get more context.

change looks intentional to fix #33321

[508d0d2] BUG: Ensure series/frame mode() keeps int index (#38732)

[galipremsagar]

@simonjayhawkins Thanks for clarifying, closing this issue.