Sample() for random rows in Data Analysis Python

Introduction

We use sample() to pick random rows from a table. This helps us look at a small, random part of big data.

You want to check a few random records from a large dataset to understand its structure.

You need to create a smaller dataset for testing or training a model.

You want to randomly select data points for a quick quality check.

You want to split data randomly for experiments or validation.

You want to shuffle data rows before analysis.

Syntax

Data Analysis Python

DataFrame.sample(n=None, frac=None, replace=False, random_state=None)

n is the number of rows to pick randomly.

frac is the fraction of rows to pick (like 0.1 for 10%).

Examples

Pick 3 random rows from the DataFrame df.

Data Analysis Python

df.sample(n=3)

Pick 20% random rows from df.

Data Analysis Python

df.sample(frac=0.2)

Pick 5 rows randomly with replacement, so rows can repeat.

Data Analysis Python

df.sample(n=5, replace=True)

Pick 4 random rows but always the same ones each time you run (for reproducibility).

Data Analysis Python

df.sample(n=4, random_state=42)

Sample Program

This code creates a small table of names and ages. Then it picks 2 random rows from it. Using random_state=1 makes sure the same rows are picked every time you run it.

Data Analysis Python

import pandas as pd

# Create a simple DataFrame
data = {'Name': ['Anna', 'Bob', 'Cara', 'Dan', 'Eva'],
        'Age': [23, 35, 45, 29, 41]}
df = pd.DataFrame(data)

# Pick 2 random rows
sampled_rows = df.sample(n=2, random_state=1)

print(sampled_rows)

Important Notes

If you use frac, do not use n at the same time.

Setting random_state helps get the same random rows every time, useful for sharing results.

By default, replace=False means rows won't repeat in the sample.

Summary

sample() helps pick random rows from data.

You can choose how many rows or what fraction to pick.

Use random_state to get repeatable random samples.

Practice

1. What does the sample() function do when used on a DataFrame in pandas?

easy

A. It groups the DataFrame by a column.

B. It sorts the DataFrame by index.

C. It selects random rows from the DataFrame.

D. It removes duplicate rows from the DataFrame.

Solution

1. Step 1: Understand the purpose of sample()

   The sample() function is designed to pick random rows from a DataFrame.

2. Step 2: Compare with other options

   Sorting, removing duplicates, and grouping are different operations and not related to sample().

3. Final Answer:

   It selects random rows from the DataFrame. -> Option C

4. Quick Check:

   sample() = random rows [OK]

Hint: Remember: sample() picks random rows only [OK]

Common Mistakes:

• Confusing sample() with sort()
• Thinking sample() removes duplicates
• Assuming sample() groups data

2. Which of the following is the correct syntax to get 5 random rows from a DataFrame named df?

easy

A. df.sample(n=5)

B. df.sample(frac=5)

C. df.sample(rows=5)

D. df.sample(size=5)

Solution

1. Step 1: Recall sample() parameters

   The correct parameter to specify number of rows is n, so df.sample(n=5) is correct.

2. Step 2: Check other options

   df.sample(frac=5) is invalid because frac must be between 0 and 1; rows and size are invalid parameters.

3. Final Answer:

   df.sample(n=5) -> Option A

4. Quick Check:

   Use n= to specify row count [OK]

Hint: Use n=5 to specify number of rows in sample() [OK]

Common Mistakes:

• Using invalid parameters like rows or size
• Forgetting to use n= when specifying count
• Using frac with count value instead of fraction

3. Given the DataFrame df below:

import pandas as pd
df = pd.DataFrame({'A': [10, 20, 30, 40, 50]})
sample_df = df.sample(n=2, random_state=1)

What will sample_df contain?

medium

A. Rows with index 1 and 4

B. Rows with index 2 and 4

C. Rows with index 0 and 3

D. Rows with index 0 and 1

Solution

1. Step 1: Understand random_state effect

   Setting random_state=1 makes sampling reproducible, so the same rows are picked every time.

2. Step 2: Check which rows are picked

   Running the code picks rows with index 1 and 4 (values 20 and 50).

3. Final Answer:

   Rows with index 1 and 4 -> Option A

4. Quick Check:

   random_state=1 fixes sample rows [OK]

Hint: Use random_state to get same sample every time [OK]

Common Mistakes:

• Ignoring random_state and expecting different rows
• Assuming sample picks first rows always
• Confusing index with row position

4. What is wrong with this code snippet to get 3 random rows from df?

sampled = df.sample(3, random_state=42, frac=0.5)

medium

A. The sample size must be greater than the DataFrame length.

B. You cannot use both n and frac together.

C. random_state must be an integer greater than 100.

D. The code is correct and will run without errors.

Solution

1. Step 1: Understand sample() parameters

   You can specify either n (number of rows) or frac (fraction of rows), but not both at the same time.

2. Step 2: Identify the error

   Using both 3 (positional for n) and frac=0.5 causes an error.

3. Final Answer:

   You cannot use both n and frac together. -> Option B

4. Quick Check:

   Use either n or frac, not both [OK]

Hint: Use only n or frac in sample(), never both [OK]

Common Mistakes:

• Trying to use n and frac together
• Thinking random_state must be >100
• Assuming sample size can exceed DataFrame length

5. You have a DataFrame df with 1000 rows. You want to randomly select 10% of the rows but ensure the sample is the same every time you run the code. Which code snippet correctly does this?

hard

A. df.sample(n=0.1, random_state=42)

B. df.sample(n=100)

C. df.sample(frac=10, random_state=42)

D. df.sample(frac=0.1, random_state=42)

Solution

1. Step 1: Understand fraction sampling

   To select 10% of rows, use frac=0.1. Using n=100 selects fixed rows (not percentage-based) and df.sample(n=100) lacks random_state so not reproducible every time.

2. Step 2: Check random_state for reproducibility

   Setting random_state=42 ensures the sample is the same every run.

3. Step 3: Identify incorrect options

   frac=10 is invalid (fraction >1), and n=0.1 is invalid because n must be integer.

4. Final Answer:

   df.sample(frac=0.1, random_state=42) -> Option D

5. Quick Check:

   Use frac=0.1 and random_state for fixed 10% sample [OK]

Hint: Use frac=0.1 and random_state for fixed 10% sample [OK]

Common Mistakes:

• Using frac > 1 instead of fraction
• Passing float to n instead of int
• Not setting random_state for reproducibility