CODEWITHSUNDEEP
pythondatetimepandasdataframedate-range
pottolom
pottolom

Reputation: 289

Python start date_range from a specific hour

I am trying to specify a date range in Python that begins on the day before the current day. However, I would like that date range to begin at 10:00:00.

This is the code I am currently using

import pandas as pd
import datetime as dt
date = dt.datetime.today() - dt.timedelta(days=1)
date_range = pd.date_range(date, freq='60min', periods=24)

However, this begins at 00:00:00. I have tried a few ways of amending the above code to get it to start at 10:00:00, but none work. Can anyone help?

To clarify: I am using Pandas because this date range is to be used as the index for a dataframe.

Upvotes: 5

Views: 3591

Answers (3)

ListenSoftware Louise Ai Agent
ListenSoftware Louise Ai Agent

Reputation: 4253

Create a string date then parse it into a datetime date

 from datetime import timedelta
 from datetime import date
 import datetime as datetime
 previous_day=pd.datetime.today().date()-datetime.timedelta(days=1)
 start_datetime=datetime.datetime.strptime(str(previous_day)+' 10:00:00','%Y-%m-%d %H:%M:%S')
 date_range = pd.date_range(start_datetime, freq='60min', periods=24)
 print(date_range)

output:

 DatetimeIndex(['2021-03-25 10:00:00', '2021-03-25 11:00:00',
                '2021-03-25 12:00:00', '2021-03-25 13:00:00',
                '2021-03-25 14:00:00', '2021-03-25 15:00:00',
                '2021-03-25 16:00:00', '2021-03-25 17:00:00',
                '2021-03-25 18:00:00', '2021-03-25 19:00:00',
                '2021-03-25 20:00:00', '2021-03-25 21:00:00',
                '2021-03-25 22:00:00', '2021-03-25 23:00:00',
                '2021-03-26 00:00:00', '2021-03-26 01:00:00',
                '2021-03-26 02:00:00', '2021-03-26 03:00:00',
                '2021-03-26 04:00:00', '2021-03-26 05:00:00',
                '2021-03-26 06:00:00', '2021-03-26 07:00:00',
                '2021-03-26 08:00:00', '2021-03-26 09:00:00'],
               dtype='datetime64[ns]', freq='60T')

Upvotes: 0

EdChum
EdChum

Reputation: 394319

You can construct another datetime but use just the day, month, year components and offset again:

In [87]:
date = dt.datetime.today() - dt.timedelta(days=1)
date = dt.datetime(date.year, date.month, date.day) + dt.timedelta(hours=10)
date_range = pd.date_range(date, freq='60min', periods=24)
date_range

Out[87]:
DatetimeIndex(['2016-11-22 10:00:00', '2016-11-22 11:00:00',
               '2016-11-22 12:00:00', '2016-11-22 13:00:00',
               '2016-11-22 14:00:00', '2016-11-22 15:00:00',
               '2016-11-22 16:00:00', '2016-11-22 17:00:00',
               '2016-11-22 18:00:00', '2016-11-22 19:00:00',
               '2016-11-22 20:00:00', '2016-11-22 21:00:00',
               '2016-11-22 22:00:00', '2016-11-22 23:00:00',
               '2016-11-23 00:00:00', '2016-11-23 01:00:00',
               '2016-11-23 02:00:00', '2016-11-23 03:00:00',
               '2016-11-23 04:00:00', '2016-11-23 05:00:00',
               '2016-11-23 06:00:00', '2016-11-23 07:00:00',
               '2016-11-23 08:00:00', '2016-11-23 09:00:00'],
              dtype='datetime64[ns]', freq='60T')

Upvotes: 5

piRSquared
piRSquared

Reputation: 294506

try

today = pd.datetime.today().date()
today - pd.offsets.Hour(14)

Timestamp('2016-11-22 10:00:00')

Then use pd.date_range

pd.date_range(today - pd.offsets.Hour(14), periods=24, freq='H')

DatetimeIndex(['2016-11-22 10:00:00', '2016-11-22 11:00:00',
               '2016-11-22 12:00:00', '2016-11-22 13:00:00',
               '2016-11-22 14:00:00', '2016-11-22 15:00:00',
               '2016-11-22 16:00:00', '2016-11-22 17:00:00',
               '2016-11-22 18:00:00', '2016-11-22 19:00:00',
               '2016-11-22 20:00:00', '2016-11-22 21:00:00',
               '2016-11-22 22:00:00', '2016-11-22 23:00:00',
               '2016-11-23 00:00:00', '2016-11-23 01:00:00',
               '2016-11-23 02:00:00', '2016-11-23 03:00:00',
               '2016-11-23 04:00:00', '2016-11-23 05:00:00',
               '2016-11-23 06:00:00', '2016-11-23 07:00:00',
               '2016-11-23 08:00:00', '2016-11-23 09:00:00'],
              dtype='datetime64[ns]', freq='H')

Upvotes: 3

Related Questions