# Your code here

tdf.rolling( '10D' ).mean()

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-49-707b0ff61efc> in <module>()
     10 tdf = df.groupby(level=1).get_group( 64413 )
     11 
---> 12 tdf.rolling( '10D' ).mean()
     13 
     14 tdf.reset_index().set_index('time').rolling( '10D' ).mean()

/home/firdaus/.conda/envs/tsconda/lib/python3.5/site-packages/pandas/core/generic.py in rolling(self, window, min_periods, freq, center, win_type, on, axis)
   5502                                    min_periods=min_periods, freq=freq,
   5503                                    center=center, win_type=win_type,
-> 5504                                    on=on, axis=axis)
   5505 
   5506         cls.rolling = rolling

/home/firdaus/.conda/envs/tsconda/lib/python3.5/site-packages/pandas/core/window.py in rolling(obj, win_type, **kwds)
   1797         return Window(obj, win_type=win_type, **kwds)
   1798 
-> 1799     return Rolling(obj, **kwds)
   1800 
   1801 

/home/firdaus/.conda/envs/tsconda/lib/python3.5/site-packages/pandas/core/window.py in __init__(self, obj, window, min_periods, freq, center, win_type, axis, on, **kwargs)
     76         self.win_type = win_type
     77         self.axis = obj._get_axis_number(axis) if axis is not None else None
---> 78         self.validate()
     79 
     80     @property

/home/firdaus/.conda/envs/tsconda/lib/python3.5/site-packages/pandas/core/window.py in validate(self)
   1055 
   1056         elif not is_integer(self.window):
-> 1057             raise ValueError("window must be an integer")
   1058         elif self.window < 0:
   1059             raise ValueError("window must be non-negative")

ValueError: window must be an integer

Problem description

The offset feature of specifying timelike windows in 'rolling' doesn't work if the dataframe has multindex with level_0 = 'time' and level_1 = something else.

Expected Output

tdf.reset_index().set_index('time').rolling( '10D' ).mean()

Works correctly.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.5.2.final.0 python-bits: 64 OS: Linux OS-release: 3.16.0-0.bpo.4-amd64 machine: x86_64 processor: byteorder: little LC_ALL: None LANG: en_US.UTF-8 LOCALE: en_US.UTF-8 pandas: 0.19.2.post+ts3 nose: None pip: 9.0.1 setuptools: 27.2.0 Cython: 0.25.2 numpy: 1.11.3 scipy: 0.18.1 statsmodels: 0.6.1 xarray: None IPython: 5.1.0 sphinx: 1.5.1 patsy: 0.4.1 dateutil: 2.6.0 pytz: 2016.10 blosc: None bottleneck: 1.2.0 tables: 3.3.0 numexpr: 2.6.1 matplotlib: 1.5.3 openpyxl: 2.4.0 xlrd: 1.0.0 xlwt: None xlsxwriter: 0.9.6 lxml: None bs4: 4.5.3 html5lib: None httplib2: None apiclient: None sqlalchemy: 1.1.4 pymysql: None psycopg2: None jinja2: 2.9.4 boto: None pandas_datareader: None

Comment From: jreback

can you show your frame construction itself.

Comment From: jreback

I suspect you are doing something like this

In [7]: df = pd.DataFrame({'value': range(60)},index=pd.MultiIndex.from_product([range(6),pd.date_range('20160101',periods=10)],names=['one','two']))

In [4]: df.rolling(window='10d').sum()
ValueError: window must be an integer

In [5]: df.rolling(window='10d',level='two').sum()

So I thought we had an issue about this, but guess not. We need to:

  • accept a level argument
  • have on interpret as per https://github.com/pandas-dev/pandas/pull/14432

Comment From: fjanoos

            time_index = pd.date_range( start, end, freq='1D',  )
            mi = pd.MultiIndex( levels=[time_index, [0]],
                             labels=[np.arange( len( time_index ) ), np.zeros( len( time_index ) )],
                             names=['time', 'int_index'] )

Comment From: hipoglucido

I am still getting this issue in version 1.5.2. Are there any plans to fix this? Thanks in advance