rolling agg raise error

[ytcpub]

Code Sample, a copy-pastable example if possible

df = pd.DataFrame(
{'code':['1', '1', '1', '3'], 'desc':['hello', 'my name is', 'rock', 'bye']},
index=pd.date_range('2018-01-01', periods=4, freq='1s')
)
df.groupby(['code']).rolling('2s').agg({'desc': 'count'})

df.groupby(['code']).rolling('2s').agg({'desc': ' '.join})
df.groupby(['code']).agg({'desc': ' '.join})

#### Problem description
i want a function like rolling with time and group concat string rows
but when use rolling('2s').agg({'desc': ' '.join}) , it raise error, I reference https://github.com/pandas-dev/pandas/issues/20773, but that code not work, not the value but direct raise exception: TypeError: cannot handle this type -> object

error log tail:
d:\program\python37\lib\site-packages\pandas\core\window.py in apply(self, func, raw, args, kwargs)
   1701     def apply(self, func, raw=None, args=(), kwargs={}):
   1702         return super(Rolling, self).apply(
-> 1703             func, raw=raw, args=args, kwargs=kwargs)
   1704
   1705     @Substitution(name='rolling')

d:\program\python37\lib\site-packages\pandas\core\window.py in apply(self, func, raw, args, kwargs)
   1010
   1011         return self._apply(f, func, args=args, kwargs=kwargs,
-> 1012                            center=False, raw=raw)
   1013
   1014     def sum(self, *args, **kwargs):

d:\program\python37\lib\site-packages\pandas\core\window.py in _apply(self, func, name, window, center, check_minp, **kwargs)
    839         results = []
    840         for b in blocks:
--> 841             values = self._prep_values(b.values)
    842
    843             if values.size == 0:

d:\program\python37\lib\site-packages\pandas\core\window.py in _prep_values(self, values, kill_inf)
    212             except (ValueError, TypeError):
    213                 raise TypeError("cannot handle this type -> {0}"
--> 214                                 "".format(values.dtype))
    215
    216         if kill_inf:

TypeError: cannot handle this type -> object

#### Expected Output
                    code        desc
2018-01-01 00:00:00    1       hello
2018-01-01 00:00:01    1  hello my name is 
2018-01-01 00:00:02    1       my name is  rock
2018-01-01 00:00:03    3         bye

#### Output of ``pd.show_versions()``

INSTALLED VERSIONS
------------------
commit: None
python: 3.7.2.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 158 Stepping 12, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.24.2
pytest: None
pip: 19.0.3
setuptools: 40.6.2
Cython: None
numpy: 1.16.2
scipy: 1.2.1
pyarrow: None
xarray: None
IPython: 7.3.0
sphinx: None
patsy: None
dateutil: 2.8.0
pytz: 2018.9
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 3.0.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml.etree: 4.3.2
bs4: 4.7.1
html5lib: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.10.1
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
gcsfs: None

[mroeschke]

As you mention, this is a duplicate of #20773.

Always open to PRs for this feature. For interim solutions, please consult Stack Overflow.