How to convert strings to date times in pandas DataFrame?

Cast strings to datetime in Pandas

You can use either the pd.datetime() or the Series.astype() to cast a DataFrame column from object to date object. Here’s how to make it work in Python:

pd.to_datetime (your_df['col_name'], format = 'your_date_format')

# or alternatively:
your_df['col_name'].astype ('datetime64[ns]', 'your_date_format')

In today’s tutorial we will learn how to change the format of object (string) columns in pandas DataFrame to the Datetime64 format.

Create Example data

import pandas as pd

dates = ['01-01-24', '01-01-24', '01-01-24', '01-01-24', '02-01-24', '02-01-24']
sales = [345, 481, 331, 387, 244, 402]

revenue = pd.DataFrame(dict (dates = dates, sales=sales))
revenue.head()

Let’s look at the data:

	dates	sales
0	01-01-24	345
1	01-01-24	481
2	01-01-24	331
3	01-01-24	387
4	02-01-24	244
5	02-01-24	402

If we look at the pandas data types of our DataFrame object we’ll see that the dates column is an object.

revenue.dtypes

This will return:

dates    object
sales     int64
dtype: object

Convert strings to dates in pandas

We have two simple ways to cast our data to the datimetime formats:

1. Using the to_datetime function:

revenue['dates'] = pd.to_datetime(revenue['dates'],format =  '%d-%m-%y')

2. Using the astype() Series method:

revenue['dates'] = revenue['dates'].astype('datetime64[ns]',  '%d-%m-%y')

Let’s look into our DataFrame:

print ( revenue.head())

Both will render the same result:

	dates	sales
0	2024-01-01	345
1	2024-01-01	481
2	2024-01-01	331
3	2024-01-01	387
4	2024-02-01	244
5	2024-02-01	402

If we check the data types we’ll see that the dates column has now a datetime type.

dates    datetime64[ns]
sales             int64
dtype: object

You can use the same technique to transform multiple DataFrame columns to dates.

Cast strings with hour, minutes and seconds to datetime

Using similar logic, you can also handle strings that has a timestamp format, and convert them to datetimes.

stamps = ['01-01-24 06:00:00', '01-01-24 12:00:00', '01-01-24 18:00:00', '02-01-24 06:00:00', '02-01-24 12:00:00', '02-01-24 12:00:00']
sales = [345, 481, 331, 387, 244, 402]
revenue = pd.DataFrame(dict (stamps = stamps, sales=sales))

We can convert the stamps column to datetime as following:

pd.to_datetime(revenue['stamps'],format =  '%d-%m-%y %H:%M:%S')

Note the format parameter, which needs to be specified to facilitate the conversion.

Now that the stamps column is a date, we can easily access its DateTimeProperties using the dt accessor, for example:

revenue['stamps'].dt.year
revenue['stamps'].dt.month_name()
revenue['stamps'].dt.isocalendar().week

Additional Learning

How to find the difference between multiple datetimes in pandas and Python?