Pandas: Select columns by data type of a given DataFrame

Last update on May 01 2025 12:54:24 (UTC/GMT +8 hours)

66. Select Columns by Data Type

Write a Pandas program to select columns by data type of a given DataFrame.

Sample Solution :

Python Code :

import pandas as pd
df = pd.DataFrame({
    'name': ['Alberto Franco','Gino Mcneill','Ryan Parkes', 'Eesha Hinton', 'Syed Wharton'],
    'date_of_birth': ['17/05/2002','16/02/1999','25/09/1998','11/05/2002','15/09/1997'],
    'age': [18.5, 21.2, 22.5, 22, 23]
})

print("Original DataFrame")
print(df)
print("\nSelect numerical columns")
print(df.select_dtypes(include = "number"))
print("\nSelect string columns")
print(df.select_dtypes(include = "object"))

Sample Output:

Original DataFrame
             name date_of_birth   age
0  Alberto Franco    17/05/2002  18.5
1    Gino Mcneill    16/02/1999  21.2
2     Ryan Parkes    25/09/1998  22.5
3    Eesha Hinton    11/05/2002  22.0
4    Syed Wharton    15/09/1997  23.0

Select numerical columns
    age
0  18.5
1  21.2
2  22.5
3  22.0
4  23.0

Select string columns
             name date_of_birth
0  Alberto Franco    17/05/2002
1    Gino Mcneill    16/02/1999
2     Ryan Parkes    25/09/1998
3    Eesha Hinton    11/05/2002
4    Syed Wharton    15/09/1997

For more Practice: Solve these Related Problems:

Write a Pandas program to select only the numeric columns from a DataFrame and then compute summary statistics on them.
Write a Pandas program to extract all object-type columns and then convert them to categorical variables.
Write a Pandas program to filter columns by data type (e.g., select datetime columns) and then output their names.
Write a Pandas program to segregate columns based on their data types and then create separate DataFrames for each type.

Go to:

Previous: Write a Pandas program to reverse order (rows, columns) of a given DataFrame.
Next: Write a Pandas program to split a given DataFrame into two random subsets.

Python-Pandas Code Editor:

Have another way to solve this solution? Contribute your code (and comments) through Disqus.