5 Best Ways to Separate Date and Time from a DateTime Column in Python Pandas

by

💡 Problem Formulation: When working with datasets in Python, often a datetime column combines both date and time information in a single field. For various analytical tasks, there’s a need to split this column into separate date and time entities. For instance, given a Pandas DataFrame with a datetime column ‘2023-03-15 08:30:00’, the goal is to have two new columns, one holding the date ‘2023-03-15′ and another for the time ’08:30:00’.

Method 1: Using dt accessor

The dt accessor in Pandas allows for easy access to date and time components of a datetime-like series. This method involves using the dt.date and dt.time attributes to extract the respective date and time components directly in a simple and straightforward manner.

Here’s an example:

import pandas as pd

# Sample DataFrame with a datetime column
df = pd.DataFrame({'datetime': pd.to_datetime(['2023-03-15 08:30:00'])})

# Separating date and time
df['date'] = df['datetime'].dt.date
df['time'] = df['datetime'].dt.time

print(df)

Output:

             datetime        date      time
0 2023-03-15 08:30:00  2023-03-15  08:30:00

This snippet creates a sample DataFrame and uses the dt accessor to assign new columns ‘date’ and ‘time’ by extracting these components from the ‘datetime’ column, effectively separating them.

Method 2: Using astype method

The astype method in Pandas can be used to convert the datetime objects to strings and then separate the date and time by string operations. This is useful when further string manipulation is needed, or when converting to a specific string format is the final goal.

Here’s an example:

import pandas as pd

# Sample DataFrame
df = pd.DataFrame({'datetime': pd.to_datetime(['2023-03-15 08:30:00'])})

# Using astype to convert to string and split
df['date'] = df['datetime'].astype(str).str.split(' ').str[0]
df['time'] = df['datetime'].astype(str).str.split(' ').str[1]

print(df)

Output:

             datetime        date      time
0 2023-03-15 08:30:00  2023-03-15  08:30:00

This code snippet converts the ‘datetime’ column to a string then splits the string on space, which separates the date and time. It then assigns the first part to the ‘date’ column and the second part to the ‘time’ column.

Method 3: Using lambda functions

For more customized splitting, lambda functions offer a flexible approach to apply any logic for separating date and time. This method is highly adaptable and great for cases that may require more complex operations than simple string or datetime accesses.

Here’s an example:

import pandas as pd

# Sample DataFrame
df = pd.DataFrame({'datetime': pd.to_datetime(['2023-03-15 08:30:00'])})

# Using a lambda function to split date and time
df['date'] = df['datetime'].apply(lambda x: x.date())
df['time'] = df['datetime'].apply(lambda x: x.time())

print(df)

Output:

             datetime        date      time
0 2023-03-15 08:30:00  2023-03-15  08:30:00

The lambda function here is applied to each element of the ‘datetime’ column, where x.date() and x.time() are called to extract the date and time, respectively. These are then set as new columns in the DataFrame.

Method 4: Using floor and subtract methods

This approach leverages the floor method to isolate the date by setting the time component to midnight. Then, the original datetime is subtracted by this floored datetime to obtain just the time. It’s a more mathematical approach and can be particularly efficient.

Here’s an example:

import pandas as pd

# Sample DataFrame
df = pd.DataFrame({'datetime': pd.to_datetime(['2023-03-15 08:30:00'])})

# Using floor and subtract to get date and time
df['date'] = df['datetime'].dt.floor('D')
df['time'] = df['datetime'] - df['date']

print(df)

Output:

             datetime       date     time
0 2023-03-15 08:30:00 2023-03-15 08:30:00

Here, df['datetime'].dt.floor('D') sets the time part of datetime to ’00:00:00′, essentially extracting the date. The original datetime minus this floor value yields a Timedelta object that represents the time.

Bonus One-Liner Method 5: List Comprehension

Using list comprehension is a Pythonic and concise way to apply any operation to a list (or a Pandas series), here separating date and time components in a single, readable line of code.

Here’s an example:

import pandas as pd

# Sample DataFrame
df = pd.DataFrame({'datetime': pd.to_datetime(['2023-03-15 08:30:00'])})

# List comprehension to separate date and time
df['date'], df['time'] = zip(*[(d.date(), d.time()) for d in df['datetime']])

print(df)

Output:

             datetime        date      time
0 2023-03-15 08:30:00  2023-03-15  08:30:00

A list of tuples is created within the list comprehension, where each tuple contains the date and time of a datetime object. The zip function then unpacks these tuples into two lists which are assigned to new DataFrame columns.

Summary/Discussion

  • Method 1: Using dt accessor. Straightforward and efficient for datetimes. Limited to datetime-like objects.
  • Method 2: Using astype method. Convenient for string operations. Might incur overhead due to type conversion.
  • Method 3: Using lambda functions. Versatile and customizable. Potentially less readable and slower for large datasets.
  • Method 4: Using floor and subtract methods. Mathematically elegant. Gives time as Timedelta rather than a time object.
  • Bonus One-Liner: List Comprehension. Pythonic and concise. Readability may decrease with complexity.