Structured arrays vs DataFrames in NumPy - Practice Questions

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of accessing structured array fields

What is the output of this code snippet that uses a NumPy structured array?

NumPy

import numpy as np

arr = np.array([(1, 2.5, 'A'), (2, 3.5, 'B')], dtype=[('x', 'i4'), ('y', 'f4'), ('z', 'U1')])
print(arr['y'])

A [1 2]

B [2.5 3.5]

C ['A' 'B']

D [1. 2.]

💡 Hint

Remember that accessing a field by name returns all values in that field.

The structured array has fields 'x', 'y', and 'z'. Accessing arr['y'] returns the array of floats [2.5, 3.5].

❓ data_output

intermediate

DataFrame column selection output

Given this pandas DataFrame, what is the output of selecting the 'age' column?

NumPy

import pandas as pd

df = pd.DataFrame({'name': ['Alice', 'Bob'], 'age': [25, 30]})
print(df['age'])

A

name    Alice
age        25
Name: 0, dtype: object

B ['Alice', 'Bob']

C [25, 30]

D

0    25
1    30
Name: age, dtype: int64

💡 Hint

Selecting a column returns a Series with index and name.

df['age'] returns a pandas Series with index 0 and 1, values 25 and 30, and name 'age'.

🧠 Conceptual

advanced

Difference in data mutability between structured arrays and DataFrames

Which statement correctly describes a key difference in mutability between NumPy structured arrays and pandas DataFrames?

A DataFrames allow adding and removing columns easily, while structured arrays have fixed fields and cannot add or remove columns.

B Structured arrays allow changing the data type of a field after creation, but DataFrames do not.

C Both structured arrays and DataFrames allow dynamic resizing of rows but not columns.

D Structured arrays support missing data natively, but DataFrames do not.

💡 Hint

Think about how flexible the structure is after creation.

NumPy structured arrays have fixed fields defined at creation and cannot add or remove columns. Pandas DataFrames can add or remove columns dynamically.

🔧 Debug

advanced

Identify the error in structured array field assignment

What error will this code produce and why?

NumPy

import numpy as np

arr = np.array([(1, 2.5), (2, 3.5)], dtype=[('a', 'i4'), ('b', 'f4')])
arr['a'] = [10, 20, 30]

A No error, assignment succeeds

B TypeError: unsupported operand type(s) for assignment

C ValueError: could not broadcast input array from shape (3,) into shape (2,)

D IndexError: index 3 is out of bounds for axis 0 with size 2

💡 Hint

Check the length of the array being assigned compared to the structured array size.

The structured array has 2 elements, but the assignment tries to assign 3 values, causing a ValueError due to shape mismatch.

🚀 Application

expert

Choosing the best data structure for mixed data types with missing values

You have a dataset with numeric, string, and missing values. You want to perform data analysis and easily handle missing data. Which data structure is best and why?

A Pandas DataFrame, because it supports mixed types and has built-in missing data handling.

B NumPy ndarray with object dtype, because it is faster and supports missing values.

C NumPy structured array, because it supports mixed types and missing values natively.

D Python list of tuples, because it is simple and supports mixed types and missing values.

💡 Hint

Consider ease of handling missing data and mixed types in analysis.

Pandas DataFrames are designed to handle mixed data types and missing values efficiently with many built-in functions.