Tutorial: How to Drop Multiple Columns in a Pandas DataFrame

[Rajeev Bagra]

🎯 Learning Objective

In this tutorial, you will learn:

✔ How to remove more than one column in Pandas
✔ How inplace=True works
✔ How to avoid errors when columns are missing
✔ Best practices for real-world datasets

---

📌 Step 1: Understanding the Context

In real-world datasets, we often create temporary columns during data cleaning.

Example:

• "lat-lon" → Used for splitting latitude and longitude
• "place_with_parent_names" → Used for location parsing

After processing, these columns are no longer needed.

Keeping them:

❌ Increases dataset size
❌ Makes analysis confusing
❌ Slows down ML pipelines

So, we remove them.

---

📌 Step 2: Sample Dataset

Let’s start with a simple example:

import pandas as pd

df1 = pd.DataFrame({
    "id": [1, 2, 3],
    "lat-lon": ["19.07,72.87", "28.70,77.10", "13.08,80.27"],
    "place_with_parent_names": ["Mumbai, India", "Delhi, India", "Chennai, India"],
    "price_usd": [120000, 150000, 90000]
})

print(df1)

Output:

id	lat-lon	place_with_parent_names	price_usd
1	19.07,72.87	Mumbai, India	120000
2	28.70,77.10	Delhi, India	150000
3	13.08,80.27	Chennai, India	90000

---

📌 Step 3: Dropping Multiple Columns

To remove more than one column, use a list:

df1.drop(
    columns=["lat-lon", "place_with_parent_names"],
    inplace=True
)

---

📌 Step 4: How This Code Works

1️⃣ columns=[...]

This tells Pandas which columns to remove:

["lat-lon", "place_with_parent_names"]

You can add more names if needed.

---

2️⃣ inplace=True

This modifies the original DataFrame.

Without it:

df1 = df1.drop(columns=[...])

With it:

df1.drop(columns=[...], inplace=True)

No reassignment needed.

---

📌 Step 5: Result After Dropping

print(df1)

Output:

id	price_usd
1	120000
2	150000
3	90000

Now the dataset is clean and focused.

---

📌 Step 6: Avoiding Errors (Best Practice)

Sometimes columns may be missing.

This causes:

KeyError

✅ Safe Version

df1.drop(
    columns=["lat-lon", "place_with_parent_names"],
    inplace=True,
    errors="ignore"
)

This skips missing columns and prevents crashes.

---

📌 Step 7: Using a Variable (Professional Style)

For large projects, use a variable:

cols_to_drop = [
    "lat-lon",
    "place_with_parent_names"
]

df1.drop(columns=cols_to_drop, inplace=True, errors="ignore")

✔ Easier to update
✔ Cleaner code
✔ Better readability

---

📌 Step 8: Verify Your Changes

Always confirm:

print(df1.columns)

Output:

Index(['id', 'price_usd'], dtype='object')

---

⭐ Step 9: Real-World Use Case

In data science projects, this method is used after:

✔ Feature extraction
✔ Data cleaning
✔ Location processing
✔ Currency conversion
✔ Text parsing

Example pipeline:

Raw Data → Clean → Extract → Drop → Analyze → Model

Dropping unused columns is a key step.

---

✅ Summary

Task	Code
Drop multiple columns	df.drop(columns=[...])
Modify original	inplace=True
Avoid errors	errors="ignore"
Best practice	Use a variable

---

📌 Final Recommended Code

cols_to_drop = ["lat-lon", "place_with_parent_names"]

df1.drop(
    columns=cols_to_drop,
    inplace=True,
    errors="ignore"
)

---

 

[Author]

Posts