Dates

The three types

datetime gives you date (a calendar day), time (a clock time), and datetime (both):

Python

from datetime import date, time, datetime

print(date(2026, 3, 14))
print(time(9, 30))
print(datetime(2026, 3, 14, 9, 30))

Output

2026-03-14
09:30:00
2026-03-14 09:30:00

Printing gives ISO format, which sorts correctly as text — a genuinely useful property.

Now

Python

from datetime import datetime, date

now = datetime.now()
print(type(now).__name__)
print(date.today().year > 2000)

Output

datetime
True

Reading the parts

Python

from datetime import datetime

moment = datetime(2026, 3, 14, 9, 30, 45)
print(moment.year, moment.month, moment.day)
print(moment.hour, moment.minute)
print(moment.weekday())
print(moment.date())

Output

weekday() counts from Monday as 0, so 5 is Saturday. isoweekday() counts from Monday as 1, which is the other convention you will meet.

Formatting with strftime

Python

from datetime import datetime

moment = datetime(2026, 3, 14, 9, 5)
print(moment.strftime("%d/%m/%Y"))
print(moment.strftime("%A %d %B %Y"))
print(moment.strftime("%H:%M"))
print(moment.strftime("%Y-%m-%d %H:%M:%S"))

Output

14/03/2026
Saturday 14 March 2026
09:05
2026-03-14 09:05:00

Code	Means
`%Y`	four-digit year
`%m`	month as 01 to 12
`%d`	day as 01 to 31
`%H`	hour, 24-hour
`%M`	minute
`%B`	month name
`%A`	weekday name

Parsing with strptime

The same codes, in reverse:

Python

from datetime import datetime

moment = datetime.strptime("14/03/2026", "%d/%m/%Y")
print(moment)
print(moment.year)

Output

2026-03-14 00:00:00
2026

The format string has to match the text exactly, character for character. A mismatch raises ValueError rather than guessing.

For ISO text, there is a shortcut that needs no format string:

Python

from datetime import datetime, date

print(datetime.fromisoformat("2026-03-14T09:30:00"))
print(date.fromisoformat("2026-03-14"))
print(date(2026, 3, 14).isoformat())

Output

2026-03-14 09:30:00
2026-03-14
2026-03-14

Differences

Subtracting two dates gives a timedelta:

Python

from datetime import date, timedelta

start = date(2026, 3, 1)
end = date(2026, 3, 14)
gap = end - start

print(gap)
print(gap.days)
print(start + timedelta(days=30))
print(start - timedelta(weeks=2))

Output

13 days, 0:00:00
13
2026-03-31
2026-02-15

timedelta takes days, weeks, hours, minutes and seconds — but not months or years, because those are not fixed lengths. For "one month later", use the dateutil package.

Comparing

Python

from datetime import date

print(date(2026, 1, 1) < date(2026, 6, 1))
print(sorted([date(2026, 5, 1), date(2026, 2, 1)]))

Output

True
[datetime.date(2026, 2, 1), datetime.date(2026, 5, 1)]

Note the list shows the repr, which is why it looks different from a printed date.

Time zones

A datetime with no zone is naive: it does not know where it is. Comparing a naive one with an aware one raises:

Python

from datetime import datetime, timezone, timedelta

aware = datetime(2026, 3, 14, 9, 0, tzinfo=timezone.utc)
print(aware)
print(aware.astimezone(timezone(timedelta(hours=2))))

naive = datetime(2026, 3, 14, 9, 0)
try:
    print(naive < aware)
except TypeError as problem:
    print("TypeError:", problem)

Output

2026-03-14 09:00:00+00:00
2026-03-14 11:00:00+02:00
TypeError: can't compare offset-naive and offset-aware datetimes

The three types

Now

Reading the parts

Formatting with strftime

Parsing with strptime

Differences

Comparing

Time zones

Test yourself

Why can a timedelta not take months or years?

What happens when you compare a naive datetime with an aware one?

Math