"can't compare datetime.date to str" in slicing/indexing (version 0.11.0)

[vfilimonov]

Hello all

I don't know if this has been already fixed in developing 0.12, but just in case:

it seemed that indexing with datetime.date does not work properly.

E.g. the code

IDX=[datetime.strptime(yy,'%Y-%m-%d') for yy in ["2005-01-01","2006-01-01","2006-01-01","2007-01-01"]]
IDX
df = pd.DataFrame([1,2,3,4], index=IDX)
print type(df.index[0])
print df.ix['2005-01-01']
df[:'2006-01-02']

works correctly. And the code

IDX=[datetime.date(datetime.strptime(yy,'%Y-%m-%d')) for yy in ["2005-01-01","2006-01-01","2006-01-01","2007-01-01"]]
IDX
df = pd.DataFrame([1,2,3,4], index=IDX)
print type(df.index[0])
print df.ix['2005-01-01']
df[:'2006-01-02']

failed with the issue "TypeError: can't compare datetime.date to str"

[jreback]

datetime.date types are not real friendly, use datetime.datetime

[vfilimonov]

OK, understood. Thank you for the reply. And are there any elegant way of converting column/index from datetime.date to datetime.datetime in DataFrame? I use pandas.io.sql.read_frame to get the table from database and types are assigned automatically.

[jreback]

it really shouldn't give you back that type anyhow (there are a couple of open bugs related to that)

try:

import pandas as pd

df['A'] = pd.to_datetime(df['A'])

you want datetimes to have the dtype datetime64[ns]

In [26]: df = DataFrame([Timestamp('20000101')]*3,index=range(3))

In [27]: df
Out[27]: 
                    0
0 2000-01-01 00:00:00
1 2000-01-01 00:00:00
2 2000-01-01 00:00:00

In [28]: df.dtypes
Out[28]: 
0    datetime64[ns]
dtype: object

[vfilimonov]

jreback, thanks a lot! pd.to_datetime was exactly what I was looking for.

[jreback]

great...i have noted the other issue (that date is not well behanved (esp from sql)).....so closing this unless you have further issues