I have tried several ways of using to_datetime, but so far I can only get it to return the dtype as "object"
pd.to_datetime(pd.Series(['28Dec2013 19:23:15']),dayfirst=True)
The return from this command is:
0 28Dec2013 19:23:15
dtype: object
You can pass a format parameter to the to_datetime function.
>>> import pandas as pd
>>> df = pd.to_datetime(pd.Series(['28Dec2013 19:23:15']),format="%d%b%Y %H:%M:%S",dayfirst=True)
>>> df
0 2013-12-28 19:23:15
dtype: datetime64[ns]
In case you need to convert existing columns in a dataframe here the solution using a helper function conv and the apply method.
import datetime
import pandas as pd
def conv(x):
return datetime.datetime.strptime(x, '%d%b%Y %H:%M:%S')
series = pd.Series(['28Dec2013 19:23:15'])
converted = series.apply(conv)
0 2013-12-28 19:23:15
dtype: datetime64[ns]
Pandas does not recognize that datetime format.
>>> pd.to_datetime(Series(['28Dec2013 19:23:15']))
0 28Dec2013 19:23:15
dtype: object
>>> pd.to_datetime(Series(['28 Dec 2013 19:23:15']))
0 2013-12-28 19:23:15
dtype: datetime64[ns]
You will need to parse the strings you are feeding into the Series. Regular expressions will likely be a good solution for this.
