I have a column of dates in this undesirable string format of 'Sun May 1 00:00:10 2016', representing 'Weekday, Mon, Day, hh:mm:ss, Year'.
How can I can format this using pandas.to_datetime or numpy?
Appreciate the help!
import pandas as pd
date = 'Sun May 1 00:00:10 2016'
df = pd.DataFrame([date], columns=['date'])
In general, for this type of tasks you should use strptime method from datetime module:
from datetime import datetime as dt
s = 'Sun May 1 00:00:10 2016'
f = '%a %B %d %H:%M:%S %Y'
dt.strptime(s, f)
datetime.datetime(2016, 5, 1, 0, 0, 10)
As far as pandas is concerned:
df = pd.DataFrame({"date":
['Sun May 1 00:00:10 2016','Sun May 2 00:00:10 2016']})
df
date
0 Sun May 1 00:00:10 2016
1 Sun May 2 00:00:10 2016
df.date.apply(lambda s: dt.strptime(s,f))
0 2016-05-01 00:00:10
1 2016-05-02 00:00:10
Name: date, dtype: datetime64[ns]
EDIT
To make it more complete, the way suggested by user1753919 in comment to your question is also "just works" in this case:
pd.to_datetime(df.date)
0 2016-05-01 00:00:10
1 2016-05-02 00:00:10
Name: date, dtype: datetime64[ns]
However, the results of timing might be of interest:
%timeit df.date.apply(lambda s: dt.strptime(s,f))
%timeit pd.to_datetime(df.date)
1000 loops, best of 3: 369 µs per loop
1000 loops, best of 3: 771 µs per loop
