Accessing fields by name in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Accessing a single field by name in a structured NumPy array

What is the output of this code snippet?

NumPy

import numpy as np

arr = np.array([(1, 2.0), (3, 4.0)], dtype=[('x', 'i4'), ('y', 'f4')])
result = arr['y']
print(result)

A [1 3]

B [2. 4.]

C [(1, 2.) (3, 4.)]

D [2 4]

💡 Hint

Accessing a field by name returns an array of that field's values.

The code creates a structured array with fields 'x' and 'y'. Accessing arr['y'] returns the values of the 'y' field as a float array.

❓ data_output

intermediate

Shape of field extracted from structured array

Given this structured array, what is the shape of the extracted field 'b'?

NumPy

import numpy as np

arr = np.array([(1, (2, 3)), (4, (5, 6))], dtype=[('a', 'i4'), ('b', 'i4', (2,))])
field_b = arr['b']
print(field_b.shape)

A (2, 2)

B (2,)

C (2, 1)

D (1, 2)

💡 Hint

The field 'b' is a 2-element integer array per record, and there are 2 records.

Each record has a field 'b' which is an array of shape (2,). Since there are 2 records, the extracted field has shape (2, 2).

🔧 Debug

advanced

Identify the error when accessing a non-existent field

What error does this code raise?

NumPy

import numpy as np

arr = np.array([(1, 2)], dtype=[('x', 'i4'), ('y', 'i4')])
print(arr['z'])

A IndexError: index out of range

B TypeError: unhashable type: 'slice'

C AttributeError: 'numpy.ndarray' object has no attribute 'z'

D KeyError: 'z'

💡 Hint

Accessing a field by a name not in the dtype raises a KeyError.

The structured array does not have a field named 'z', so accessing arr['z'] raises a KeyError.

🚀 Application

advanced

Extracting multiple fields and combining them

Given this structured array, which option correctly creates a new array combining fields 'name' and 'age' as a list of tuples?

NumPy

import numpy as np

arr = np.array([('Alice', 25), ('Bob', 30)], dtype=[('name', 'U10'), ('age', 'i4')])

A np.array((arr['name'], arr['age']))

B arr[['name', 'age']].tolist()

C list(zip(arr['name'], arr['age']))

D [arr['name'], arr['age']]

💡 Hint

Use zip to pair elements from two arrays into tuples.

Using zip on arr['name'] and arr['age'] pairs each name with its age, then list converts it to a list of tuples.

🧠 Conceptual

expert

Understanding memory layout when accessing fields by name

When you access a field by name in a structured NumPy array, what best describes the memory layout of the returned array?

A It is a view of the original array sharing the same memory.

B It is a new array with data copied and reshaped.

C It is a deep copy with independent memory.

D It is a list of Python objects extracted from the array.

💡 Hint

Accessing a field returns a view, not a copy, unless explicitly copied.

Accessing a field by name returns a view into the original array's memory, so changes to the field affect the original array.