rolling window function with multiple arguments by group

[robpla]

Code Sample, a copy-pastable example if possible

## let's say we have the below DF
rand = np.random.RandomState(1)
dff = pd.DataFrame({'A' : np.arange(20),
                    'B' : rand.randint(100, 120, 20),
                    'C' : rand.randint(0, 2, 20)})

### Ideal solution, this does not work of course
#dff['tau'] = dff.groupby('C').sort_values('A').rolling(window = 5).apply(sp.stats.mstats.kendalltau, args = ({'x' : 'A', 'y' : 'B'))

## workaround - need to pass an index, use a global DF, and (?) hard-code columns to use, use a to-be-deprecated function etc etc
def my_tau_indx(indx):
    x = dff.iloc[indx, 0]
    y = dff.iloc[indx, 1]
    tau = sp.stats.mstats.kendalltau(x, y)[0]
    return tau

grp = dff.sort_values(['A', 'C']).groupby('C', group_keys=False)
func = lambda x: pd.Series(pd.rolling_apply(np.arange(len(x)), 5, my_tau_indx), x.index)
t = grp.apply(func)
dff.reindex(t.index).assign(tau=t)

Problem description

There is no simple way to run a scipy/custom function requiring multiple arguments (by group) in a rolling window.

In the code above, let's say that the 'C' column below is used for grouping.
I need to

• Group by column 'C'
• Within each group, sort by 'A'
• Withing each group, apply a rolling function taking two (or more) arguments, like kendalltau, to arguments 'A' and 'B'.

I asked this question here: http://stackoverflow.com/questions/41715814/python-pandas-rolling-function-with-two-arguments-in-a-grouped-dataframe/41719105#41719105 .

And the only viable solution I received uses a to-be-deprecated pd.rolling_apply.

Expected Output

See the link above for the expected results. See the code sample for what I'd like to write to get the result.

Output of pd.show_versions()

# Paste the output here pd.show_versions() here

DEBUG:pip.vcs:Registered VCS backend: git
DEBUG:pip.vcs:Registered VCS backend: hg
DEBUG:pip.pep425tags:Config variable 'Py_DEBUG' is unset, Python ABI tag may be incorrect
DEBUG:pip.pep425tags:Config variable 'WITH_PYMALLOC' is unset, Python ABI tag may be incorrect
DEBUG:pip.pep425tags:Config variable 'Py_DEBUG' is unset, Python ABI tag may be incorrect
DEBUG:pip.pep425tags:Config variable 'WITH_PYMALLOC' is unset, Python ABI tag may be incorrect
DEBUG:pip.vcs:Registered VCS backend: svn
DEBUG:pip.vcs:Registered VCS backend: bzr

INSTALLED VERSIONS

commit: None
python: 3.5.2.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 62 Stepping 4, GenuineIntel
byteorder: little
LC_ALL: None
LANG: en
LOCALE: None.None

pandas: 0.19.2
nose: None
pip: 9.0.1
setuptools: 27.2.0
Cython: None
numpy: 1.12.0rc2
scipy: 0.18.1
statsmodels: None
xarray: None
IPython: 5.1.0
sphinx: 1.4.8
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2016.7
blosc: None
bottleneck: None
tables: None
numexpr: None
matplotlib: 1.5.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.8
boto: None
pandas_datareader: None

[TomAugspurger]

Dupe of #15095

Still trying to figure out the API. We're potentially constrained by backwards-compat. Post there if you have any thought.