Python - Subset DataFrame by Column Name

Using Pandas library, we can perform multiple operations on a DataFrame. We can even create and access the subset of a DataFrame in multiple formats. The task here is to create a subset DataFrame by column name. We can choose different methods to perform this task. Here are possible methods mentioned below -

Before performing any action, we need to write few lines of code to import necessary libraries and create a DataFrame.

Creating the DataFrame

#import pandas
import pandas as pd

# create dataframe
data = {'Name': ['John', 'Emily', 'Lara', 'Lucas', 'Katy', 'Jordan'], 
        'Gender': [30, 27, 21, 21, 16, 20], 
        'Branch': ['Arts', 'Arts', 'Commerce', 'Science', 
                   'Science', 'Science'],
        'pre_1': [9, 9, 10, 7, 6, 9],
        'pre_2': [8, 7, 10, 6, 8, 8]}

df = pd.DataFrame(data)
df

Output:

Method 1: Using Python iloc() function 

This function allows us to create a subset by choosing specific values from columns based on indexes. 

Syntax:

df_name.iloc[beg_index:end_index+1,beg_index:end_index+1]

Example: Create a subset with Name, Gender and Branch column

# create a subset of all rows
# and Name, Gender and Branch column
df.iloc[:, 0:3]

Output :

Method 2: Using Indexing Operator

We can use the indexing operator i.e. square brackets to create a subset dataframe 

Example: Create a subset with Name, pre_1, and pre_2 column

# creating subset dataframe using 
# indexing operator
df[['Name', 'pre_1', 'pre_2']]

Output - 

Method 3: Using filter() method with like keyword

We can use this method particularly when we have to create a subset dataframe with columns having similarly patterned names.

Example: Create a subset with pre_1 and pre_2 column

# create a subset of columns pre_1 and pre_2
# using filter() method
df.filter(like='pre')

Output: