python - Cast types on the fly while creating dataframe -

i have kind of partial json object, dict:

{     u'20140816 00': {u'var': u'40.78'},     u'20140816 01': {u'var': u'53.24'},     u'20140816 02': {u'var': u'50.23'},     ... }

and want put in pandas dataframe.

however, change keys string datetime (datetime.strptime(key, '%y%m%d %h') , cast var value string float.

is possible while creating pandas dataframe, or if not easiest way recast types json object pandas dataframe.

you can use convert_objects method convert strings desired dtypes.

normally read_ methods have date handling param , these can handle formats.

in case, seeing convert_objects doesn't quite want can call pandas method to_datetime convert column so:

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

O9