Get Timestamps of DatetimeIndex that are around Timestamp-CodePudding

I have a Timestamp object and a DatetimeIndex object:

Out[18]: Timestamp('2022-05-04 11:03:51.138289-0400', tz='America/Toronto')
Out[33]: DatetimeIndex(['2022-05-04 10:00:00-04:00', '2022-05-04 10:30:00-04:00',
                        '2022-05-04 11:00:00-04:00', '2022-05-04 11:30:00-04:00',
                        '2022-05-04 12:00:00-04:00', '2022-05-04 12:30:00-04:00',
                        '2022-05-04 13:00:00-04:00', '2022-05-04 13:30:00-04:00',
                        '2022-05-04 14:00:00-04:00', '2022-05-04 14:30:00-04:00',
                        '2022-05-04 15:00:00-04:00', '2022-05-04 15:30:00-04:00',
                        '2022-05-04 16:00:00-04:00'],
                       dtype='datetime64[ns, America/Toronto]', freq=None)

How can I get the Timestamp object within DatetimeIndex that is immediately before and immediately after Timestamp?

For example,

For:

Timestamp('2022-05-04 11:03:51.138289-0400', tz='America/Toronto')

Return:

DatetimeIndex(['2022-05-04 11:00:00-04:00', '2022-05-04 11:30:00-04:00'], 
              dtype='datetime64[ns, America/Toronto]', freq=None)

I can do something totally hacky like so:

import pandas
from pandas_market_calendars import get_calendar, date_range

tz = "America/Toronto"

nyse = get_calendar("NYSE")

now = pandas.Timestamp.now(tz=tz)
dti = date_range(nyse.schedule("2022-05-04", "2022-05-04"), frequency="30min").tz_convert(tz)
pre = []
post = []

print(now)

for i in dti:
    if i <= now:
        pre.append(i)
    if i >= now:
        post.append(i)

print(pre[-1])
print(post[0])

out:
2022-05-04 12:54:18.236651-04:00
2022-05-04 12:30:00-04:00
2022-05-04 13:00:00-04:00

But I feel like there are better ways to do this, be it list comprehension or some built-in pandas method that I'm unfamiliar with, or something else.

It was suggested that Converting between datetime, Timestamp and datetime64 could be an answer, but I don't believe it's appropriate. That question is around conversion, my question is around ranges.

CodePudding user response：

Assuming dti is sorted, one option is to use numpy.searchsorted because it finds the index in dti where now should be inserted to maintain order:

import numpy as np
idx = np.searchsorted(dti, now)
out = dti[idx-1:idx 1]

Output:

DatetimeIndex(['2022-05-04 11:00:00-04:00', '2022-05-04 11:30:00-04:00'], dtype='datetime64[ns, America/Toronto]', freq=None)