Record arrays in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of accessing fields in a record array

What is the output of the following code?

NumPy

import numpy as np

rec = np.rec.array([(1, 2.0, 'Hello'), (2, 3.0, 'World')],
                  dtype=[('foo', 'i4'), ('bar', 'f4'), ('baz', 'a5')])

print(rec.foo)
print(rec['baz'])

A

[1 2]
[b'Hello' b'World']

B

[1 2]
['Hello' 'World']

C

[1. 2.]
[b'Hello' b'World']

D

[1 2]
[b'Hello' b'World' b'!']

💡 Hint

Remember that record array fields keep their original data types, including byte strings.

The field 'foo' contains integers, so printing rec.foo outputs an integer array. The field 'baz' is a byte string type, so printing rec['baz'] outputs byte strings.

❓ data_output

intermediate

Number of elements in a record array after filtering

Given the record array below, how many elements remain after filtering where age > 30?

NumPy

import numpy as np

rec = np.rec.array([(25, 'Alice'), (35, 'Bob'), (40, 'Charlie')],
                  dtype=[('age', 'i4'), ('name', 'U10')])

filtered = rec[rec.age > 30]

print(len(filtered))

A 0

B 1

C 2

D 3

💡 Hint

Count how many ages are greater than 30.

Only 'Bob' (35) and 'Charlie' (40) have ages greater than 30, so 2 elements remain.

🔧 Debug

advanced

Identify the error in record array field assignment

What error does the following code raise?

NumPy

import numpy as np

rec = np.rec.array([(1, 2.0), (3, 4.0)], dtype=[('x', 'i4'), ('y', 'f4')])
rec.x = [10, 20, 30]

A ValueError: setting an array element with a sequence.

B IndexError: too many indices for array

C AttributeError: can't set attribute

D TypeError: invalid type assignment

💡 Hint

Check if the length of the assigned list matches the record array length.

The record array has length 2, but the list assigned has length 3, causing a ValueError due to shape mismatch.

🚀 Application

advanced

Create a record array from a structured array

You have a structured array arr with dtype [('a', 'i4'), ('b', 'f4')]. Which code correctly converts it to a record array?

NumPy

import numpy as np

arr = np.array([(1, 2.0), (3, 4.0)], dtype=[('a', 'i4'), ('b', 'f4')])

A rec = np.recarray(arr.shape, dtype=arr.dtype)

B rec = np.rec.array(arr)

C rec = np.array(arr, dtype=np.recarray)

D rec = arr.view(np.recarray)

💡 Hint

Use the view method to change the array type without copying data.

Using arr.view(np.recarray) creates a record array view of the structured array without copying data. Option D creates a new record array but copies data. Option D is invalid dtype usage. Option D creates an empty record array.

🧠 Conceptual

expert

Why use record arrays instead of structured arrays?

Which statement best explains the advantage of using NumPy record arrays over structured arrays?

A Record arrays support multi-dimensional fields, unlike structured arrays.

B Record arrays allow accessing fields as attributes, making code more readable and concise.

C Record arrays automatically convert all string fields to Unicode.

D Record arrays use less memory than structured arrays for the same data.

💡 Hint

Think about how you access data fields in each array type.

Record arrays allow accessing fields using dot notation (e.g., rec.field), which can make code easier to write and read. Structured arrays require dictionary-like access (e.g., arr['field']). Memory usage and string handling are similar, and both support multi-dimensional fields.