datetime gets converted to int64 in groupby agg

[twiecki]

I tried to produce a stand-alone example but couldn't so far. Maybe the answer is already apparent though. This started happening just recently (maybe after updated pandas to 0.17.1) but worked well before. I'm doing a multi-column agg in a groupby. One of the columns is a datetime of which I want the first element:

(Pdb) t.groupby(['block_dir']).first()
                                 dt    sid  amount     price symbol  \
block_dir                                                             
1         2003-01-02 15:56:00+00:00  21719     -62  0.963811   AIRN   

Works fine, however:

          order_sign  block_time  
block_dir                         
1              False           0  
(Pdb) t.groupby(['block_dir']).agg({'dt': 'first'})
                            dt
block_dir                     
1          1041522960000000000

Expected Output

output of pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.11.final.0
python-bits: 64
OS: Linux
OS-release: 3.13.0-77-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: en_US.UTF-8
LANG: en_US.UTF-8

pandas: 0.17.1
nose: 1.3.7
pip: 8.0.2
setuptools: 19.6.2
Cython: 0.23.4
numpy: 1.10.4
scipy: 0.16.1
statsmodels: 0.6.1
IPython: 3.2.1
sphinx: 1.3.1
patsy: 0.4.0
dateutil: 2.4.2
pytz: 2015.4
blosc: None
bottleneck: 1.0.0
tables: 3.2.2
numexpr: 2.4.3
matplotlib: 1.4.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.3.2
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: 1.0.8
pymysql: None
psycopg2: 2.6.1 (dt dec pq3 ext)
Jinja2: None

[jorisvandenbossche]

This reproduces it (it is the timezone that is the trigger):

In [1]: pd.__version__
Out[1]: u'0.17.1'

In [2]: df = pd.DataFrame({'col':['a','a','b','b'], 'date':pd.date_range('2015-0
1-01', periods=4, utc=True)})

In [3]: df.groupby('col').first()
Out[3]:
          date
col
a   2015-01-01
b   2015-01-03

In [4]: df.groupby('col').agg({'date': 'first'})
Out[4]:
          date
col
a   2015-01-01
b   2015-01-03

In [5]: df['date'] = df['date'].dt.tz_localize('utc')

In [6]: df.groupby('col').first()
Out[6]:
                         date
col
a   2015-01-01 00:00:00+00:00
b   2015-01-03 00:00:00+00:00

In [7]: df.groupby('col').agg({'date': 'first'})
Out[7]:
                    date
col
a    1420070400000000000
b    1420243200000000000

[twiecki]

Thanks @jorisvandenbossche.

[jorisvandenbossche]

And it seems a duplicate of #11616, fixed in master by #11672

[twiecki]

I'll close it then.