How to Drop Columns in Pandas DataFrame (With Examples)

When you’re working with data in Python using Pandas, your dataset often comes with more information than you need. Perhaps your CSV contains extra metadata, there are duplicate fields, or you simply want to simplify your analysis. There you need to drop columns from Pandas dataframe.

Think of your DataFrame like a spreadsheet with too many columns. Before you can run meaningful analysis, you want to clean it up — just like clearing unnecessary columns from an Excel sheet. In Pandas, you do this with the drop() method (and a few other handy techniques).

In this guide, we’ll walk you through every method for removing columns from a Pandas DataFrame, explain the details behind each, and share real-world scenarios where this is relevant.

Remove Columns From Pandas DataFrame

There are multiple ways to drop columns from Pandas dataframe. Let’s go through each method:

The drop() Method in Pandas

The most common way to remove columns is with the drop() method.

Syntax

DataFrame.drop(
    labels=None,
    axis=0,
    index=None,
    columns=None,
    level=None,
    inplace=False,
    errors='raise'
)

Parameters

• labels → The name(s) of rows or columns to remove.
• axis → 0 means rows, 1 means columns.
• columns → Shortcut for removing columns (recommended for clarity).
• inplace → If True, modifies the original DataFrame. If False (default), returns a new DataFrame.
• errors → By default, Pandas will throw an error if you try to drop something that doesn’t exist. Use errors="ignore" to avoid this.

For most column removal tasks, you’ll simply use:

df.drop(columns=["ColumnName"])

Drop a Single Column

Example: removing "ExtraNotes" from a student grades dataset.

import pandas as pd

data = {
    "Name": ["Ali", "Sara", "John", "Mehak"],
    "Math": [85, 92, 78, 88],
    "Science": [90, 95, 80, 85],
    "ExtraNotes": ["-", "-", "-", "-"]
}

df = pd.DataFrame(data)

print("Original DataFrame:\n", df)

# Drop a single column
df = df.drop(columns=["ExtraNotes"])

print("\nModified DataFrame:\n", df)

Output:

Original DataFrame:
     Name  Math  Science ExtraNotes
0    Ali    85       90          -
1   Sara    92       95          -
2   John    78       80          -
3  Mehak    88       85          -
Modified DataFrame:
     Name  Math  Science
0    Ali    85       90
1   Sara    92       95
2   John    78       80
3  Mehak    88       85

Best practice: use the columns= parameter instead of axis=1 for readability.

Drop Multiple Columns

Need to drop more than one? Just pass a list of column names.

# Drop 'Math' and 'Science'
df = df.drop(columns=["Math", "Science"])

This is helpful in machine learning pipelines where you want to remove multiple features at once.

Drop Columns by Index

Sometimes column names are messy or duplicated. In those cases, you can drop by index:

# Drop the first two columns by index
df = df.drop(columns=df.columns[[0, 1]])

Tip: Always check column order with df.columns before using indices, since order can shift after merges or imports.

1. Using .pop()

Removes a column and returns it as a Series.

removed = df.pop("Math")
print("Removed column:\n", removed)

Think of it like pulling a drawer out of a cabinet — you remove it, but you still keep the drawer itself.

2. Using del

Quickly deletes a column (in-place).

del df["Science"]

⚠️ Be careful: unlike drop(), there’s no undo — it modifies immediately.

3. Selecting a Subset of Columns

Instead of deleting, you can choose the columns you want to keep:

df = df[["Name", "Math"]]

This is often the cleanest approach when you know exactly which features you need.

Real-World Scenarios

Here’s where column dropping makes sense:

• Data cleaning: Removing "Unnamed: 0" columns created during CSV export.
• Text preprocessing: Dropping metadata like "comments" or "URLs".
• Machine learning prep: Excluding "ID" or "Timestamp" columns that don’t add predictive value.
• Performance optimization: Large logs with 100+ fields can be trimmed to the 10–15 you need.

Common Mistakes & Tips

• Forgetting to set axis=1 (or columns=) → Pandas defaults to rows, not columns.
• Expecting the original DataFrame to change → Use inplace=True or reassign (df = df.drop(...)).
• Dropping non-existent columns → Use errors="ignore" if you’re not sure.
• Overusing inplace=True → Sometimes keeping both original and modified versions helps avoid mistakes.

FAQs: Pandas Drop Column

Q1: What’s the difference between drop() and pop()?

• drop() → Removes multiple columns, returns DataFrame.
• pop() → Removes one column, returns Series.

Q2: Can I drop rows with drop()?
Yes — just set axis=0 or use the index= parameter.

Q3: Which is faster, drop() or del?
del is slightly faster (since it works in-place), but drop() is safer and more flexible.

Q4: How do I drop duplicate columns?
You can use:

df = df.loc[:, ~df.columns.duplicated()]

Q5: How do I drop columns with missing values?

df = df.dropna(axis=1)

This removes all columns with at least one NaN.

Conclusion

Dropping columns in Pandas is a fundamental data-cleaning skill. Whether you use drop(), pop(), del, or column selection, the goal is the same: streamline your DataFrame for cleaner analysis and faster processing.

Key takeaways:

• Use drop(columns=...) for clarity.
• Be mindful of inplace=True.
• Explore pop() if you want the removed column separately.
• Always think in terms of what data you truly need.

Ready to try it yourself? Experiment with dropping columns right now in our Online Python Compiler — practice by creating a sample DataFrame and cleaning it up.

For more interesting articles, follow our Python tutorial series by SyntaxScenarios.