df.rolling(...).sum() on DataFrame containing string columns returns a TypeError

[alporter08]

Code Sample

df = pd.DataFrame({'B': [0, 1, 2, np.nan, 4]},
                   index = [pd.Timestamp('20130101 09:00:00'),
                            pd.Timestamp('20130101 09:00:02'),
                            pd.Timestamp('20130101 09:00:03'),
                            pd.Timestamp('20130101 09:00:05'),
                            pd.Timestamp('20130101 09:00:06')])

df = df.reset_index().rename(columns={'index':'timestamp'})

df['id'] = 'abcd'


df.rolling('2s', on='timestamp').sum()

Original dataframe:

timestamp	B	id
2013-01-01 09:00:00	0.0	abcd
2013-01-01 09:00:02	1.0	abcd
2013-01-01 09:00:03	2.0	abcd
2013-01-01 09:00:05	NaN	abcd
2013-01-01 09:00:06	4.0	abcd

Problem description

Rolling used to work in prior versions (e.g. 0.20.3) on a dataframe that included string columns. When running the above code in version 0.23.4, the following error is raised:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/Users/aporter/anaconda3/lib/python3.6/site-packages/pandas/core/window.py in _prep_values(self, values, kill_inf)
    221             try:
--> 222                 values = _ensure_float64(values)
    223             except (ValueError, TypeError):

pandas/_libs/algos_common_helper.pxi in pandas._libs.algos.ensure_float64()

pandas/_libs/algos_common_helper.pxi in pandas._libs.algos.ensure_float64()

ValueError: could not convert string to float: 'abcd'

During handling of the above exception, another exception occurred:

TypeError                                 Traceback (most recent call last)
<ipython-input-11-e98474769a85> in <module>()
----> 1 df.rolling('2s', on='timestamp').sum()

/Users/aporter/anaconda3/lib/python3.6/site-packages/pandas/core/window.py in sum(self, *args, **kwargs)
   1584     def sum(self, *args, **kwargs):
   1585         nv.validate_rolling_func('sum', args, kwargs)
-> 1586         return super(Rolling, self).sum(*args, **kwargs)
   1587 
   1588     @Substitution(name='rolling')

/Users/aporter/anaconda3/lib/python3.6/site-packages/pandas/core/window.py in sum(self, *args, **kwargs)
   1005     def sum(self, *args, **kwargs):
   1006         nv.validate_window_func('sum', args, kwargs)
-> 1007         return self._apply('roll_sum', 'sum', **kwargs)
   1008 
   1009     _shared_docs['max'] = dedent("""

/Users/aporter/anaconda3/lib/python3.6/site-packages/pandas/core/window.py in _apply(self, func, name, window, center, check_minp, **kwargs)
    842         results = []
    843         for b in blocks:
--> 844             values = self._prep_values(b.values)
    845 
    846             if values.size == 0:

/Users/aporter/anaconda3/lib/python3.6/site-packages/pandas/core/window.py in _prep_values(self, values, kill_inf)
    223             except (ValueError, TypeError):
    224                 raise TypeError("cannot handle this type -> {0}"
--> 225                                 "".format(values.dtype))
    226 
    227         if kill_inf:

TypeError: cannot handle this type -> object

Expected Output

In 0.20.3, the output is:

timestamp	B	id
2013-01-01 09:00:00	0.0	abcd
2013-01-01 09:00:02	1.0	abcd
2013-01-01 09:00:03	3.0	abcd
2013-01-01 09:00:05	NaN	abcd
2013-01-01 09:00:06	4.0	abcd

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.1.final.0 python-bits: 64 OS: Darwin OS-release: 16.7.0 machine: x86_64 processor: i386 byteorder: little LC_ALL: None LANG: en_US.UTF-8 LOCALE: en_US.UTF-8

pandas: 0.23.4
pytest: 3.0.7
pip: 18.1
setuptools: 40.0.0
Cython: 0.25.2
numpy: 1.14.2
scipy: 0.19.0
pyarrow: None
xarray: None
IPython: 5.3.0
sphinx: 1.5.6
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2017.2
blosc: None
bottleneck: 1.2.1
tables: 3.3.0
numexpr: 2.6.2
feather: None
matplotlib: 2.0.2
openpyxl: 2.4.7
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.3
bs4: 4.6.0
html5lib: 0.999
sqlalchemy: 1.1.9
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[mroeschke]

Thanks for the report. Another user reported a similar issue #23002. Closing in favor of that issue, but feel free to chime in there.