indexing

[behzadnouri]

import pandas as pd
from numpy.random import rand
idx = pd.date_range(start='5/20/2013', end='5/21/2013')
df = pd.DataFrame({k:(100*rand(len(idx))).astype('int') for k in ['6', '36']}, index=idx)

everything looks fine:

>>> df
            36   6
2013-05-20  58  23
2013-05-21  68  31

and even this works:

>>> df['6']
2013-05-20    23
2013-05-21    31
Freq: D, Name: 6, dtype: int32

but this is one fails:

>>> df['36']
Empty DataFrame
Columns: [36, 6]
Index: []

it seems to me the problem comes from a 'DatetimeIndex', because this works fine:

>>> df.index = [d.date() for d in df.index]
>>> df['36']
2013-05-20    58
2013-05-21    68
Name: 36, dtype: int32

[jreback]

This was fixed in master via:
#3597

Pls confirm. Thanks for the report

In [9]: pandas.__version__
Out[9]: '0.12.0.dev-7441af9'

In [5]: df = pd.DataFrame({k:(100*rand(len(idx))).astype('int') for k in ['6', '36']}, index=idx)

In [6]: df
Out[6]: 
            36   6
2013-05-20  51  45
2013-05-21   1  13

In [7]: df['6']
Out[7]: 
2013-05-20    45
2013-05-21    13
Freq: D, Name: 6, dtype: int64

In [8]: df['36']
Out[8]: 
2013-05-20    51
2013-05-21     1
Freq: D, Name: 36, dtype: int64

[jreback]

did you try with master?