Drop columns in DataFrame by label Names or by Index Positions

A pandas DataFrame is a 2D data structure for storing tabular data. When working with DataFrames, you often need to remove unwanted columns. This can be done by specifying column names or their index positions using the drop() method.

In this tutorial, we'll explore different methods to drop columns from a pandas DataFrame including dropping by names, index positions, and ranges.

Creating the Sample DataFrame

Let's start by creating a sample DataFrame to work with ?

import pandas as pd

dataset = {
    "Employee ID": ["CIR45", "CIR12", "CIR18", "CIR50", "CIR28"], 
    "Age": [25, 28, 27, 26, 25], 
    "Salary": [200000, 250000, 180000, 300000, 280000], 
    "Role": ["Junior Developer", "Analyst", "Programmer", "Senior Developer", "HR"]
}

df = pd.DataFrame(dataset, index=["Nimesh", "Arjun", "Mohan", "Ritesh", "Raghav"])
print(df)

        Employee ID  Age  Salary              Role
Nimesh        CIR45   25  200000  Junior Developer
Arjun         CIR12   28  250000           Analyst
Mohan         CIR18   27  180000        Programmer
Ritesh        CIR50   26  300000  Senior Developer
Raghav        CIR28   25  280000                HR

Method 1: Drop Columns by Names

The most straightforward way is to specify column names directly. Use axis=1 to indicate column operations ?

import pandas as pd

dataset = {
    "Employee ID": ["CIR45", "CIR12", "CIR18", "CIR50", "CIR28"], 
    "Age": [25, 28, 27, 26, 25], 
    "Salary": [200000, 250000, 180000, 300000, 280000], 
    "Role": ["Junior Developer", "Analyst", "Programmer", "Senior Developer", "HR"]
}

df = pd.DataFrame(dataset, index=["Nimesh", "Arjun", "Mohan", "Ritesh", "Raghav"])

# Drop columns by name
result = df.drop(["Role", "Salary"], axis=1)
print("After dropping Role and Salary columns:")
print(result)

After dropping Role and Salary columns:
        Employee ID  Age
Nimesh        CIR45   25
Arjun         CIR12   28
Mohan         CIR18   27
Ritesh        CIR50   26
Raghav        CIR28   25

Method 2: Drop Columns by Index Positions

You can also drop columns using their index positions. Use df.columns[[indices]] to select columns by position ?

import pandas as pd

dataset = {
    "Employee ID": ["CIR45", "CIR12", "CIR18", "CIR50", "CIR28"], 
    "Age": [25, 28, 27, 26, 25], 
    "Salary": [200000, 250000, 180000, 300000, 280000], 
    "Role": ["Junior Developer", "Analyst", "Programmer", "Senior Developer", "HR"]
}

df = pd.DataFrame(dataset, index=["Nimesh", "Arjun", "Mohan", "Ritesh", "Raghav"])

# Drop columns by index (Salary=2, Role=3)
result = df.drop(df.columns[[2, 3]], axis=1)
print("After dropping columns at positions 2 and 3:")
print(result)

After dropping columns at positions 2 and 3:
        Employee ID  Age
Nimesh        CIR45   25
Arjun         CIR12   28
Mohan         CIR18   27
Ritesh        CIR50   26
Raghav        CIR28   25

Method 3: Drop Range of Columns Using iloc

Use iloc to select a range of columns by position. Note that iloc excludes the end index ?

import pandas as pd

dataset = {
    "Employee ID": ["CIR45", "CIR12", "CIR18", "CIR50", "CIR28"], 
    "Age": [25, 28, 27, 26, 25], 
    "Salary": [200000, 250000, 180000, 300000, 280000], 
    "Role": ["Junior Developer", "Analyst", "Programmer", "Senior Developer", "HR"]
}

df = pd.DataFrame(dataset, index=["Nimesh", "Arjun", "Mohan", "Ritesh", "Raghav"])

# Drop columns from index 1 to 4 (Age, Salary, Role)
result = df.drop(df.columns[1:4], axis=1)
print("After dropping columns from position 1 to 3:")
print(result)

After dropping columns from position 1 to 3:
        Employee ID
Nimesh        CIR45
Arjun         CIR12
Mohan         CIR18
Ritesh        CIR50
Raghav        CIR28

Method 4: Drop Range of Columns Using loc

Use loc with column names to drop a range. Unlike iloc, loc includes both start and end columns ?

import pandas as pd

dataset = {
    "Employee ID": ["CIR45", "CIR12", "CIR18", "CIR50", "CIR28"], 
    "Age": [25, 28, 27, 26, 25], 
    "Salary": [200000, 250000, 180000, 300000, 280000], 
    "Role": ["Junior Developer", "Analyst", "Programmer", "Senior Developer", "HR"]
}

df = pd.DataFrame(dataset, index=["Nimesh", "Arjun", "Mohan", "Ritesh", "Raghav"])

# Drop columns from Age to Role (inclusive)
result = df.drop(df.loc[:, "Age":"Role"].columns, axis=1)
print("After dropping columns from Age to Role:")
print(result)

After dropping columns from Age to Role:
        Employee ID
Nimesh        CIR45
Arjun         CIR12
Mohan         CIR18
Ritesh        CIR50
Raghav        CIR28

Comparison of Methods

Conclusion

Pandas provides flexible methods to drop columns from DataFrames. Use column names for clarity, index positions for programmatic operations, and loc/iloc for range-based dropping. Always specify axis=1 when dropping columns.