python - Applying datetime format in pandas for sorting -

in pandas, have column dates of format: %y/%m/%d (e.g. 2015/10/31). want change format like: %d-%m-%y (e.g. 31-10-15).

turning column correct object later sorting:

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

applying strptime:

df['date'] = df['date'].apply(lambda x:     datetime.strptime(x,'%d-%m-%y'))  typeerror: must str, not timestamp 

also, oddly, if dates enter pandas dataframe in other default iso standard weird results occur on sorting. such inconsistent formats and/or not sorting:

0  2015-01-31 1  2016-15-01 

because dealing datetime object already, getting error because strptime requires string , not timestamp object. definition of strptime:

def strptime(cls, date_string, format):         'string, format -> new datetime parsed string (like time.strptime()).' 

what looking first convert datetime format require in string using strftime:

def strftime(self, format):         """return string representing date , time, controlled         explicit format string. 

and bring datetime object using strptime. following demo demonstrate. note use of .date() @ end in order remove unneeded 00:00:00 time portion.

>>> datetime import datetime >>> orig_datetime_obj = datetime.strptime("2015/10/31", '%y/%m/%d').date() >>> print(orig_datetime_obj) 2015-10-31 >>> print(type(orig_datetime_obj)) <type 'datetime.datetime'> >>> new_datetime_obj = datetime.strptime(orig_datetime_obj.strftime('%d-%m-%y'), '%d-%m-%y').date() >>> print(new_datetime_obj) 2015-10-31 >>> print(type(new_datetime_obj)) <type 'datetime.date'> 

alternatively, if require converting different format in string, can stick using strftime new format. using example above, need portion:

orig_datetime_obj.strftime('%d-%m-%y')