How to Exclude Columns in Pandas?

Excluding columns in a Pandas DataFrame is a common operation when you want to work with only relevant data. In this article, we will discuss various methods to exclude columns from a DataFrame, including using .loc[], .drop(), and other techniques.

Exclude One Column using .loc[]

We can exclude a column by its location using the .loc [] function. The code below demonstrates how to exclude a specific column by comparing column names.

import pandas as pd

d = pd.DataFrame({'food_id': [1, 2, 3, 4],
                     'name': ['idly', 'dosa', 'poori', 'chapathi'],
                     'city': ['delhi', 'goa', 'hyd', 'chennai'],
                     'cost': [12, 34, 21, 23]})

# Exclude the 'cost' column
ex = d.loc[:, d.columns != 'cost']
print("DataFrame after excluding 'cost' column using .loc[]")
print(ex)

Output:

DataFrame after excluding 'cost' column using .loc[]

     food_id      name     city
0        1             idly         delhi
1        2            dosa        goa
2        3           poori        hyd
3        4         chapathi   chennai

If we need to exclude multiple columns, we can use .isin() function.

# exclude name and food_id column
print(d.loc[:, ~d.columns.isin(['name', 'food_id'])])

Output

      city         cost
0    delhi       12
1      goa        34
2      hyd        21
3   chennai   23

Other methods to exclude columns in Pandas, apart from using DataFrame.loc[], are discussed below:

Excluding Columns Using .remove()

Although .remove() is not a direct Pandas method, we can use it with a list of column names before reassigning the DataFrame.

# Convert columns to a list and remove the column 'cost'
columns = list(d.columns)
columns.remove('cost')

# Create a new DataFrame with the remaining columns
ex = d[columns]
print(ex)

Output

     food_id    name     city
0        1            idly       delhi
1        2           dosa      goa
2        3         poori       hyd
3        4       chapathi  chennai

Excluding Columns by Name Using drop()

drop() method is one of the most common ways to exclude specific columns by name:

# Exclude the 'cost' column
ex = d.drop(columns=['cost'])
print(ex)

Output

    food_id     name     city
0        1            idly        delhi
1        2           dosa      goa
2        3          poori      hyd
3        4       chapathi  chennai

Exclude Columns by Data Type Using select_dtypes()

We can exclude columns based on their data type using the select_dtypes() method.

# Exclude all numeric columns
ex = d.select_dtypes(exclude=['number'])
print(ex)

Output

       name     city
0      idly       delhi
1      dosa      goa
2     poori      hyd
3  chapathi  chennai

Excluding Columns Dynamically Using List Comprehensions

If we need to exclude columns dynamically based on a condition, we can use list comprehensions. For example, to exclude columns whose names start with a specific letter:

# Exclude columns whose names start with 'c'
ex = d[[col for col in d.columns if not col.startswith('c')]]
print(ex)

Output

    food_id     name
0        1           idly
1        2           dosa
2        3          poori
3        4         chapathi

These techniques make column exclusion in Pandas flexible and straightforward, helping us focus on the data we need.