pivot_table produces inconsistent columns if applied to empty table

[ischurov]

Code Sample, a copy-pastable example if possible

# Part 1
df1 = pd.DataFrame([], columns=['a', 'b', 'value'])
pivot1 = df1.pivot_table(index='a', columns='b', values='value', 
                         aggfunc='count')
print(pivot1.columns)

#Output: 
MultiIndex(levels=[['value'], []],
           labels=[[], []],
           names=[None, 'b'])

# Part 2
df2 = pd.DataFrame([[1, 2, 3]], columns=['a', 'b', 'value'])
pivot2 = df2.pivot_table(index='a', columns='b', values='value', 
                         aggfunc='count')
print(pivot2.columns)

#Output:
Int64Index([2], dtype='int64', name='b')

Problem description

In the first example, I don't expect to see any multiindex in .columns, as exactly one value for columns is provided. As we don't have any data, and it is not possible to figure out the dtype of this index, one can probably assume it's something like Index([], name='b').

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.7.0.final.0 python-bits: 64 OS: Darwin OS-release: 17.6.0 machine: x86_64 processor: i386 byteorder: little LC_ALL: en_US.UTF-8 LANG: en_US.UTF-8 LOCALE: en_US.UTF-8

pandas: 0.23.3
pytest: None
pip: 10.0.1
setuptools: 39.2.0
Cython: None
numpy: 1.14.5
scipy: 1.1.0
pyarrow: None
xarray: None
IPython: 6.4.0
sphinx: None
patsy: 0.5.0
dateutil: 2.7.3
pytz: 2018.5
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.2.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: 1.0.1
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[gfyoung]

Indeed, that does look a little odd! Perhaps a corner case?

cc @jreback

[mroeschke]

Looks to work on master now. Could use a test

In [10]: df1 = pd.DataFrame([], columns=['a', 'b', 'value'])
    ...: pivot1 = df1.pivot_table(index='a', columns='b', values='value',
    ...:                          aggfunc='count')
    ...: print(pivot1.columns)
Index([], dtype='object', name='b')

[FaazAbidi]

Is anyone working on it? I can take this. @mroeschke

[mroeschke]

Go for it @FaazAbidi

[FaazAbidi]

take

[FaazAbidi]

@mroeschke I think this test belongs to tests.frame.test_repr_info.py. Can you confirm?

[mroeschke]

Probably pandas/tests/reshape/test_pivot.py

[MirijaH]

take