BUG: groupy().nth() throws error on multiple groups, empty result

[adbull]

Code Sample, a copy-pastable example if possible

>>> import pandas as pd
>>> df = pd.DataFrame(index=[0], columns=['a', 'b', 'c'])
>>> df.groupby('a').nth(10)

Empty DataFrame
Columns: [b, c]
Index: []

>>> df.groupby(['a', 'b']).nth(10)

Traceback (most recent call last):
  File "<ipython-input-3-ae8299c3984e>", line 1, in <module>
    df.groupby(['a', 'b']).nth(10)
  File "~/anaconda3/lib/python3.5/site-packages/pandas/core/groupby.py", line 1390, in nth
    return out.sort_index() if self.sort else out
  File "~/anaconda3/lib/python3.5/site-packages/pandas/core/frame.py", line 3344, in sort_index
    indexer = lexsort_indexer(labels._get_labels_for_sorting(),
  File "~/anaconda3/lib/python3.5/site-packages/pandas/core/indexes/multi.py", line 1652, in _get_labels_for_sorting
    for label in self.labels]
  File "~/anaconda3/lib/python3.5/site-packages/pandas/core/indexes/multi.py", line 1652, in <listcomp>
    for label in self.labels]
  File "~/anaconda3/lib/python3.5/site-packages/numpy/core/_methods.py", line 26, in _amax
    return umr_maximum(a, axis, None, out, keepdims)
ValueError: zero-size array to reduction operation maximum which has no identity

Problem description

In the current Github version of Pandas, when calling groupby().nth() with multiple grouping columns, an error is raised if the result is empty. This is a regression from version 0.19.2.

Expected Output

Empty DataFrame
Columns: [b, c]
Index: []

Empty DataFrame
Columns: [c]
Index: []

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.5.2.final.0
python-bits: 64
OS: Linux
OS-release: 4.9.8-100.fc24.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: C
LANG: C
LOCALE: None.None

pandas: 0.19.0+829.gb17e286
pytest: 3.0.5
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.25.2
numpy: 1.11.3
scipy: 0.18.1
xarray: 0.9.1
IPython: 4.2.0
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: 2.6.2
feather: None
matplotlib: 2.0.0
openpyxl: 2.4.1
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.2
bs4: 4.5.3
html5lib: 0.999
sqlalchemy: 1.1.5
pymysql: None
psycopg2: None
jinja2: 2.9.4
s3fs: None
pandas_gbq: None
pandas_datareader: None

[adbull]

Think this just requires:

diff --git a/pandas/core/indexes/multi.py b/pandas/core/indexes/multi.py
index 92baf9d..34b62c5 100644
--- a/pandas/core/indexes/multi.py
+++ b/pandas/core/indexes/multi.py
@@ -1645,11 +1645,9 @@ class MultiIndex(Index):
         """
         from pandas.core.categorical import Categorical
 
-        return [Categorical.from_codes(label,
-                                       np.arange(np.array(label).max() + 1,
-                                                 dtype=label.dtype),
-                                       ordered=True)
-                for label in self.labels]
+        return [Categorical.from_codes(label, np.arange(
+            np.array(label).max() + 1 if len(label) else 0,
+            dtype=label.dtype), ordered=True) for label in self.labels]
 
     def sortlevel(self, level=0, ascending=True, sort_remaining=True):
         """

[jreback]

can u put up a PR with that fix (and test)?