ENH: Export DataFrame containing datetime with tz to HDF5 in fixed format

[Tux1]

Hello,

I noticed a bug about exporting DataFrame with columns containing localized datetime. I might be due to some another required libraries updated...
Example :

a = pd.to_datetime(['20120201', '20130201'])
b = pd.to_datetime(['20120201', '20130201'])
dfa = pd.DataFrame([a])
dfb = pd.DataFrame([b])
dfb[0] = dfb[0].dt.tz_localize('UTC')

dfa.to_hdf('tata.hdf', 'df') # success
dfb.to_hdf('toto.hdf', 'df') # fail

It works fine with Series
Thanks guys

[jreback]

you need to use format='table', this is not implemented for format='fixed', the default.

[Tux1]

However, it was working until pandas 0.17, no ?

[jreback]

It was never supported in any real way (and not tested), so it 'worked' I guess.

In [7]: dfb.to_hdf('toto.hdf', 'df') # fail
/Users/jreback/miniconda/envs/pandas_0.16.2/lib/python2.7/site-packages/pandas/io/pytables.py:2577: PerformanceWarning: 
your performance may suffer as PyTables will pickle object types that it cannot
map directly to c-types [inferred_type->datetime,key->block1_values] [items->[0]]

  warnings.warn(ws, PerformanceWarning)

[jreback]

this is a dupe of #8165

if you'd like to work on that would be great.

[Tux1]

Not very easy but I will try yes.
Thanks

[jreback]

actually, will reopen this one and close the other