to_datetime 以及 dt.days、dt.months

Series类型的数据，经过 to_datetime 之后就可以用 pandas.Series.dt.day 和 pandas.Series.pd.month。

import pandas as pd
a  = pd.Series(['2017-1-2','2017-9-1'])
print(a)

0    2017-1-2
1    2017-9-1
dtype: object

a = pd.to_datetime(a)
print(a)

0   2017-01-02
1   2017-09-01
dtype: datetime64[ns]

b = pd.Series(['2019-8-19', '2019-6-29'])
b = pd.to_datetime(b)
print(b)

0   2019-08-19
1   2019-06-29
dtype: datetime64[ns]

c = (b-a).dt.days
print(c)

0    959
1    666
dtype: int64

import datetime
d = datetime.date(2019,8,19) - a.dt.date
print(d)

0   959 days
1   717 days
dtype: timedelta64[ns]

d = (datetime.date(2019,8,19) - a.dt.date).dt.days
print(d)

0    959
1    717
dtype: int64