printing a dataframe ignores timezone information of columns containing datetime64

[satizabal]

I have a dataframe with a column containing dates (I do not want to use this column as the index of my dataframe). When printing the dataframe, the timezone information is lost. Directly accessing the .values shows the correct timezone.

This might be related to #6415

import pandas as pd

print "pandas version : ", pd.__version__

series = pd.to_datetime([1426374000000, 1426415697429], unit='ms').tz_localize('UTC').tz_convert('Europe/Berlin')
print '-- Printing series --'
print series
df = pd.DataFrame(series)
print '-- Printing df --'
print df
print '-- Printing df.values --'
print df.values

pandas version :  0.16.0
-- Printing series --
<class 'pandas.tseries.index.DatetimeIndex'>
[2015-03-15 00:00:00+01:00, 2015-03-15 11:34:57.429000+01:00]
Length: 2, Freq: None, Timezone: Europe/Berlin
-- Printing df --
                           0
0        2015-03-14 23:00:00
1 2015-03-15 10:34:57.429000
-- Printing df.values --
[['2015-03-15T00:00:00.000000000+0100']
 ['2015-03-15T11:34:57.429000000+0100']]

[jreback]

this will be resolved by #8260

currently the dtype of datetimes with timezones are of object dtype (and NOT datetime64[ns]
and hence there is not timezone information indicated on the column. It IS on the individual objects.

So this is a highly inefficient way to do things, but was simpler when it was implemented (quite a while ago).