I have spent over a decade working with data in Python, and if there is one thing I have learned, it is that data is rarely perfect when you first load it.
Usually, I find myself staring at a “Price” column that is somehow a string or a “Date” column that Python thinks is just a bunch of numbers.
In this tutorial, I will show you how to change the data type of a column in Pandas using the same methods I use in my daily production work.
Use the astype() Method
The astype() method is my go-to tool when I need to perform a quick and explicit conversion of a column.
It is easy and allows you to cast a pandas object to a specified dtype very efficiently.
In the example below, I have a dataset of tech employees in California where the ‘Salary’ was imported as a string, and I need it to be a float for calculations.
import pandas as pd
# Creating a dataset of tech employees in San Francisco
data = {
'Employee_Name': ['Alice', 'Bob', 'Charlie', 'David'],
'City': ['San Francisco', 'Los Angeles', 'San Diego', 'San Jose'],
'Salary': ['125000.50', '98000.00', '110000.75', '135000.25'],
'Years_Exp': ['5', '3', '7', '10']
}
df = pd.DataFrame(data)
# Checking original data types
print("Original Dtypes:")
print(df.dtypes)
# Changing 'Salary' to float and 'Years_Exp' to integer
df['Salary'] = df['Salary'].astype(float)
df['Years_Exp'] = df['Years_Exp'].astype(int)
print("\nUpdated Dtypes:")
print(df.dtypes)You can refer to the screenshot below to see the output.
One thing I really like about astype() is that you can pass a dictionary to change multiple columns at once, which keeps the code clean.
Change Multiple Columns with a Dictionary
When I am dealing with large DataFrames, I prefer not to write a new line for every single column conversion.
Using a dictionary with astype() is much cleaner and easier to maintain in a professional codebase.
import pandas as pd
# Data representing retail store inventory in Chicago
inventory_data = {
'Store_ID': [101, 102, 103],
'Product': ['Laptop', 'Monitor', 'Keyboard'],
'Price': ['1200.50', '300.00', '45.99'],
'Quantity': ['15', '40', '100']
}
df = pd.DataFrame(inventory_data)
# Using a dictionary to change multiple types at once
df = df.astype({
'Price': float,
'Quantity': int
})
print(df.dtypes)You can refer to the screenshot below to see the output.
Use pd.to_numeric for Robust Conversions
Sometimes, astype() fails because the data contains “dirty” values like a stray “N/A” string or a random character.
In these cases, I always use pd.to_numeric() because it has an errors parameter that is a lifesaver.
Setting errors=’coerce’ will turn any non-convertible value into a NaN, preventing your whole script from crashing.
import pandas as pd
# Data from a New York housing survey with some messy entries
housing_data = {
'Apartment_ID': ['A1', 'B2', 'C3', 'D4'],
'Square_Feet': ['1200', '950', 'MISSING', '1100'],
'Monthly_Rent': ['3200', '2800', '3500', 'invalid_data']
}
df = pd.DataFrame(housing_data)
# Converting to numeric and handling errors gracefully
df['Square_Feet'] = pd.to_numeric(df['Square_Feet'], errors='coerce')
df['Monthly_Rent'] = pd.to_numeric(df['Monthly_Rent'], errors='coerce')
print(df)
print("\nColumn Types:")
print(df.dtypes)You can refer to the screenshot below to see the output.
Convert Strings to Datetime
Working with dates is notorious for being difficult, but pd.to_datetime() makes it significantly easier.
I use this constantly when analyzing American holiday sales or stock market trends where dates come in various formats.
import pandas as pd
# Sales data for a 4th of July promotion
sales_data = {
'Sale_Date': ['2024-07-01', '2024-07-02', '2024-07-03', '2024-07-04'],
'Revenue': [4500, 5200, 6100, 8900]
}
df = pd.DataFrame(sales_data)
# Changing 'Sale_Date' from object to datetime64
df['Sale_Date'] = pd.to_datetime(df['Sale_Date'])
print(df.dtypes)
# Now we can easily extract the day of the week
df['Day_of_Week'] = df['Sale_Date'].dt.day_name()
print(df[['Sale_Date', 'Day_of_Week']])The convert_dtypes() Method
If you are feeling a bit lazy or have a massive DataFrame and want Pandas to take its best guess at the types, use convert_dtypes().
This was introduced in newer versions of Pandas to automatically convert columns to the best possible “nullable” types.
import pandas as pd
# Sample dataset with mixed types
data = {
'A': [1, 2, 3],
'B': ['a', 'b', 'c'],
'C': [True, False, True],
'D': [1.5, 2.5, 3.5]
}
df = pd.DataFrame(data)
# Force everything to object type first to simulate a raw import
df = df.astype(object)
# Let Pandas automatically infer and convert to best types
df_optimized = df.convert_dtypes()
print(df_optimized.dtypes)In this tutorial, we looked at several ways to change the data type of a column in Pandas.
Choosing the right method depends on whether you have clean data or need to handle missing and “dirty” values.
You may also read:
- How to Check if a Pandas DataFrame is Empty
- Ways to Get the First Row of a Pandas DataFrame
- How to Get Column Names in Pandas
- How to Check Pandas Version in Python
I am Bijay Kumar, a Microsoft MVP in SharePoint. Apart from SharePoint, I started working on Python, Machine learning, and artificial intelligence for the last 5 years. During this time I got expertise in various Python libraries also like Tkinter, Pandas, NumPy, Turtle, Django, Matplotlib, Tensorflow, Scipy, Scikit-Learn, etc… for various clients in the United States, Canada, the United Kingdom, Australia, New Zealand, etc. Check out my profile.
