Keeping first vs last vs none in Pandas - When to Use Which

The Big Idea

What if you could clean messy data with one simple command instead of hours of tedious work?

The Scenario

Imagine you have a big list of customer orders with some customers appearing multiple times. You want to find unique customers by keeping only their first order, last order, or removing all duplicates entirely.

The Problem

Doing this by hand means scanning the list over and over, comparing each entry, and deciding which to keep. This is slow, confusing, and easy to make mistakes, especially with large data.

The Solution

Using pandas' drop_duplicates with the 'keep' option lets you quickly choose to keep the first, last, or no duplicates. It handles all the hard work efficiently and correctly.

Before vs After

✗ Before

unique_orders = []
for order in orders:
  if order.customer not in [o.customer for o in unique_orders]:
    unique_orders.append(order)

✓ After

df.drop_duplicates(subset='customer', keep='first')  # or 'last' or False

What It Enables

This lets you easily clean and prepare data for analysis, focusing on exactly the records you need without errors or wasted time.

Real Life Example

A sales analyst wants to see only the first purchase date per customer to study buying patterns. Using 'keep=first' quickly filters the data to just those records.

Key Takeaways

Manually removing duplicates is slow and error-prone.

pandas drop_duplicates with 'keep' option automates this task.

You can choose to keep first, last, or no duplicates easily.

Practice

1.

In pandas, what does keep='first' do when removing duplicates?

easy

A. Keeps the first occurrence of each duplicate and removes the rest

B. Keeps the last occurrence of each duplicate and removes the rest

C. Removes all duplicates, keeping only unique rows

D. Keeps all duplicates without removing any

Solution

1. Step 1: Understand the meaning of keep='first'

   This option keeps the first row of each set of duplicates and removes the later ones.

2. Step 2: Compare with other options

   Unlike keep='last' which keeps the last occurrence, keep='first' keeps the earliest one.

3. Final Answer:

   Keeps the first occurrence of each duplicate and removes the rest -> Option A

4. Quick Check:

   keep='first' = Keep first duplicate [OK]

Hint: Remember: 'first' keeps earliest duplicate row [OK]

Common Mistakes:

• Confusing keep='first' with keep='last'
• Thinking keep='first' removes all duplicates
• Assuming keep='first' keeps all duplicates

2.

Which of the following is the correct syntax to remove duplicates in a pandas DataFrame df keeping only the last occurrence?

df._____('column_name', keep='last')

easy

A. drop_dupes

B. remove_duplicates

C. drop_duplicates

D. delete_duplicates

Solution

1. Step 1: Identify the pandas method for removing duplicates

   The correct method is drop_duplicates to remove duplicate rows.

2. Step 2: Confirm syntax for keeping last occurrence

   Using keep='last' inside drop_duplicates keeps the last duplicate row.

3. Final Answer:

   drop_duplicates -> Option C

4. Quick Check:

   Method to remove duplicates = drop_duplicates [OK]

Hint: Use df.drop_duplicates() to remove duplicates [OK]

Common Mistakes:

• Using non-existent methods like remove_duplicates
• Confusing method names with similar words
• Forgetting to specify keep='last'

3.

What will be the output of the following code?

import pandas as pd

data = {'A': [1, 2, 2, 3, 3, 3], 'B': ['x', 'y', 'y', 'z', 'z', 'z']}
df = pd.DataFrame(data)
result = df.drop_duplicates(keep=False)
print(result)

medium

A. Rows with unique values only, no duplicates

B. All rows except the first occurrence of duplicates

C. Rows with A=1 B='x', A=2 B='y', A=3 B='z' only

D. Empty DataFrame

Solution

1. Step 1: Understand keep=False behavior

   Using keep=False removes all duplicates, keeping only rows that appear once.

2. Step 2: Identify unique rows in the DataFrame

   In the data, only the row with A=1 and B='x' is unique; others appear multiple times.

3. Final Answer:

   Rows with unique values only, no duplicates -> Option A

4. Quick Check:

   keep=False = Keep only unique rows [OK]

Hint: keep=False drops all duplicates, keeps only unique rows [OK]

Common Mistakes:

• Thinking keep=False keeps first or last duplicates
• Expecting some duplicates to remain
• Confusing output with keep='first' or keep='last'

4.

Consider this code snippet:

import pandas as pd

data = {'col': [1, 1, 2, 2, 3]}
df = pd.DataFrame(data)
result = df.drop_duplicates(keep='middle')
print(result)

What is the error in this code?

medium

A. DataFrame missing required columns

B. Invalid value for the 'keep' parameter

C. Syntax error in DataFrame creation

D. No error, code runs fine

Solution

1. Step 1: Check the 'keep' parameter values

   The valid values for 'keep' are 'first', 'last', or False. 'middle' is invalid.

2. Step 2: Identify the error raised by invalid 'keep'

   Using 'middle' causes a ValueError because pandas does not recognize it.

3. Final Answer:

   Invalid value for the 'keep' parameter -> Option B

4. Quick Check:

   keep must be 'first', 'last', or False [OK]

Hint: Use only 'first', 'last', or False for keep parameter [OK]

Common Mistakes:

• Using unsupported strings for keep
• Assuming 'middle' is valid
• Ignoring error messages

5.

You have a DataFrame df with duplicate rows based on column 'ID'. You want to remove all duplicates completely, so only rows with unique 'ID' remain. Which code snippet achieves this?

1) df.drop_duplicates(subset=['ID'], keep='first')
2) df.drop_duplicates(subset=['ID'], keep='last')
3) df.drop_duplicates(subset=['ID'], keep=False)
4) df.drop_duplicates()

hard

A. Option 1

B. Option 4

C. Option 2

D. Option 3

Solution

1. Step 1: Understand the goal

   The goal is to remove all duplicates so only unique 'ID' rows remain.

2. Step 2: Analyze each option

   Options 1 and 2 keep one duplicate (first or last). Option 4 removes duplicates but keeps first by default. Only option 3 with keep=False removes all duplicates completely.

3. Final Answer:

   Option 3 -> Option D

4. Quick Check:

   keep=False removes all duplicates [OK]

Hint: Use keep=False to remove all duplicates fully [OK]

Common Mistakes:

• Using keep='first' or 'last' when full removal needed
• Not specifying subset for duplicate check
• Assuming default drop_duplicates removes all duplicates