熊猫根据给定（任意）的datetimeindex重新采样（例如使用最近的值）

Question

Is it possible to use pandas.resample with a given (arbitrary) DatetimeIndex (using eg nearest option with a given time window), instead of a the rule string for regular dates?

EDIT:

Example:

dates = pd.DatetimeIndex(['2000-01-01 12:00:00', '2000-01-03 13:00:00', '2000-01-05 15:00:00', '2000-01-09 10:00:00'])
df = pd.DataFrame({'dummy': dates}, index=dates)
custom_dates = pd.DatetimeIndex(['2000-01-02 09:00:00', '2000-01-05 22:00:00', '2000-01-10 15:00:00'])
new_df = df.resample(custom_dates, method='nearest')

And new_df should now have as DatetimeIndex custom_dates and the columns from df .

Answer 1

Maybe a bit late, but here is a solution that uses reindex which supports your desired nearest option:

import pandas as pd

dates = pd.DatetimeIndex(['2000-01-01 12:00:00', '2000-01-03 13:00:00', '2000-01-05 15:00:00', '2000-01-09 10:00:00'])
df = pd.DataFrame({'dummy': dates}, index=dates)
custom_dates = pd.DatetimeIndex(['2000-01-02 09:00:00', '2000-01-05 22:00:00', '2000-01-10 15:00:00'])

df.reindex(custom_dates, method='nearest', tolerance=pd.Timedelta(2, 'D'))

Output:

                    dummy
2000-01-02 09:00:00 2000-01-01 12:00:00
2000-01-05 22:00:00 2000-01-05 15:00:00
2000-01-10 15:00:00 2000-01-09 10:00:00