Reference guide: Datetime manipulation

The following tables can serve as reference guides to remind you of the shorthand code for
manipulating datetime strings into individual objects.

Manipulating datetime strings in Python

Below, you will find a table with the datetime functions you can use to help you manipulate
datetime objects in different ways.

Code Format Example

%a Abbreviated workday Sun

%A Weekday Sunday

%b Abbreviated month Jan

%B Month name January

%c Date and time Sun Jan 1 00:00:00 2021

%d Day (leading zeros) 01 to 31

%H 24 hours 00 to 23

%I 12 hours 01 to 12

%j Day of year 001 to 366

%m Month 01 to 12

%M Minute 00 to 59

%p AM or PM AM/PM

%S Seconds 00 to 61

%U Week number (Sun) 00 to 53

%W Week number (Mon) 00 to 53

%w Weekday 0 to 6

%x Locale’s appropriate date 08/16/88 (None);

representation 08/16/1988 (en_US);
16.08.1988 (de_DE)
%X A locale’s appropriate time 21:30:00 (en_US);
representation 21:30:00 (de_DE)

%y Year without century 00 to 99

%Y Year 2022

%z Offset +0900

%Z Time zone EDT/JST/WET etc (GMT)

 Datetime functions to remember
All of the following date string manipulations require the datetime package to be imported first.

Code Input Type Input Example Output Output

Type Example

datetime.strptime(“25/11/2022”, string “25/11/2022” DateTime “2022-11-25

“%d/%m/%Y”) 00:00:00”

datetime.strftime(dt_object, DateTime “2022-11-25 string “25/11/2022”

“%d/%m/%Y”) 00:00:00”

dt_object = string “25/11/2022” float (UTC 1617836400.0

datetime.strptime(“25/11/2022”, timestamp
“%d/%m/%Y”) in
datetime.timestamp(dt_object) seconds)

datetime.strptime(“25/11/2022”, string “25/11/2022” string “2022-11-25”

“%d/%m/%Y”).strftime(“%Y-%m-%d”)

datetime.fromtimestamp(1617836400. float 1617836400.0 DateTime datetime.date

0) (UTC time(2021, 4,
timestamp 7, 23, 0)
in
seconds)

datetime.fromtimestamp(1617836400. float 1617836400.0 string “'07/04/2021'

0).strftime(“%d/%m/%Y”) (UTC ”
timestamp
in
seconds)

from pytz import timezone string NewYork DateTime Tokyo

ny_time = timezone timezone
datetime.strptime(“25-11-2022 “25-11-2022 2022, 11, 26,
09:34:00-0700”, “%d-%m-%Y 09:34:00-0700 1, 34,
%H:%M:%S%z”) ” JST+9:00:00
Tokyo_time = STD>
ny_time.astimezone(timezone(‘Asia/
Tokyo’))

datetime.strptime(“20:00”, string “20:00” string “08:00 PM”

“%H:%M”).strftime(“%I:%M %p”)

datetime.strptime(“08:00 PM”, string “08:00 PM” string “20:00”

“%I:%M %p”).strftime(“%H:%M”)
Datetime in NumPy and pandas
A preface regarding terminology in the following section: datetime refers to the specific
module of that name in the Python standard library or to the specific class within that module.
Datetime (or uncapitalized, datetime) refers to any date/time-related object from any library or
language.

You’ve learned that the datetime module in Python’s standard library contains a number of
classes used to work with time data, including date, time, datetime, timedelta,
timezone, and tzinfo. Remember, modules are similar to libraries, in that they are groups of
related classes and functions, but they are generally subcomponents of libraries. Classes are
data types that bundle data and functionality together.

NumPy and pandas have their own datetime classes that offer significant performance boosts
when working with large datasets. Pandas datetime classes, like the rest of the pandas library,
are built on NumPy. These classes have very similar (and in many cases identical) functionality
to Python’s native datetime classes, but they run more efficiently due to NumPy and pandas’
vectorization capabilities. Therefore, although you can use datetime data in pandas, it’s
generally better to use NumPy or pandas datetime objects when working in pandas, if possible.

NumPy’s datetime classes include, most notably, datetime64 and timedelta64. Like
datetime objects, datetime64 objects contain date and time information in a single data
structure; and, like timedelta objects, timedelta64 objects contain information pertaining
to spans of time.

Pandas’ datetime classes include Timestamp, Timedelta, Period, and DateOffset.

Because these classes are efficient and dynamic in their capabilities, you often don’t need to
import the datetime module when working with datetime data in pandas. Also, pandas will
automatically recognize datetime-like data and convert it to the appropriate class when
possible. Here’s an example:

data = ['2023-01-20', '2023-04-27', '2023-06-15']

my_series = pd.Series(data)
my_series
0 2023-01-20
1 2023-04-27
2 2023-06-15
dtype: object

This series contains string data, but it can be converted to datetime64 data using the
pd.to_datetime() function:
my_series = pd.to_datetime(my_series)
my_series

0 2023-01-20
1 2023-04-27
2 2023-06-15
dtype: datetime64[ns]

Refer to the pandas to_datetime() documentation for more information about this function.
When a Series object contains datetime data, you can use dt to access various properties of
the data. For example:
print(my_series.dt.year)
print()
print(my_series.dt.month)
print()
print(my_series.dt.day)

0 2023
1 2023
2 2023
dtype: int64

0 1
1 4
2 6
dtype: int64

0 20
1 27
2 15
dtype: int64
Note that it’s not uncommon to import the datetime module from Python’s standard library as
dt. You may have encountered this yourself. In such case, dt is being used as an alias. The
pandas dt Series accessor (as demonstrated in the last example) is a different thing entirely.
Refer to the pandas dt accessor documentation for more information.

Key takeaways

Use reference guides like the tables above throughout your career to help remind you of the
different ways to manipulate datetime objects. Even experts in the field use reference guides,
rather than memorizing all this information. Getting familiar with guides like these will be
beneficial because you will be using them throughout your career as a data professional.