Creating a range of dates in Python

Control over your dates with datetime and timedelta

One-liner solution for date range

Opt for Python's built-in datetime and timedelta if you need more customization:

from datetime import datetime, timedelta

def generate_dates(start_date, end_date):
    return [(start_date + timedelta(days=i)).isoformat() for i in range((end_date - start_date).days + 1)]

start_date = datetime.strptime('2021-01-01', '%Y-%m-%d')
end_date = datetime.strptime('2021-01-10', '%Y-%m-%d')
dates_list = generate_dates(start_date, end_date)
print(dates_list)


Here, list comprehension works hard to push dates into our list, one day at a time.

Pythonic way for memory efficiency

If the range is too big, a generator keeps the memory use minimal:

def date_generator(start_date, end_date):
    current_date = start_date
    while current_date <= end_date:
        yield current_date
        current_date += timedelta(days=1)

for single_date in date_generator(start_date, end_date):
    print(single_date.isoformat())


Now, we're saving memory while we time travel. 🚀

Taming large ranges with islice

itertools.islice can handle gargantuan date ranges without breaking a sweat:

import itertools

large_range = itertools.islice(date_generator(start_date, end_date), 1000)


This will efficiently handle the first 1000 dates only, and memory won't even notice it.

Optimizing date range creation

Using Python's dateutil for complex ranges

The python-dateutil package has got your back for complex date recurrences:

from dateutil.rrule import rrule, DAILY

dates_range = list(rrule(DAILY, dtstart=start_date, until=end_date))


'cause sometimes we need to be complex, right? 🧙‍♂️

Convert pandas Timestamps to Python datetimes

To convert pandas Timestamps to Pythonic datetime objects, «voilà!»:

py_dates = [d.to_pydatetime() for d in pd.date_range(start_date, end_date)]


Converted them just like that! 🔁

Time zones? pandas got your back

Working with date ranges across time zones? No problem, pandas has got you covered:

tz_dates = pd.date_range(start_date, periods=10, tz='UTC')
print(tz_dates)


Yeah, cross-timezone relationships are no more complicated! ⌚✈️

Fun with NumPy: dates and speed

NumPy's datetime for speed

If you fancy NumPy and need speed, use datetime64:

import numpy as np

np_dates = np.arange('2021-01-01', '2021-01-11', dtype='datetime64[D]')
print(np_dates)


There you go, 0 to 60 dates in no time! 🏁