BinGrouper breaks on DataFrame.transform

[dalejung]

import pandas as pd
import numpy as np
from pandas.tseries.resample import TimeGrouper
import pandas.util.testing as tm

df = tm.makeTimeDataFrame()
tg = TimeGrouper('M')

func = lambda x: (x - 1)
df.groupby(tg).transform(func)

/externals/pandas/pandas/core/groupby.py in _get_compressed_labels(self)
   1360     def _get_compressed_labels(self):
-> 1361         all_labels = [ping.labels for ping in self.groupings]
   1362         if self._overflow_possible:

TypeError: 'NoneType' object is not iterable

Seems related to #8049 as that check group_info which errors on BinGrouper since there are no groupings

[jreback]

hmm breaking on the heuristic to figure out if we have the right label set

I can take a stab if u would like
I prob will look later

[jreback]

merged in
pls try and report back