np.unique() for unique elements in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of np.unique() with return_counts

What is the output of this code snippet using np.unique() with return_counts=true?

NumPy

import numpy as np
arr = np.array([3, 1, 2, 3, 2, 1, 4])
result = np.unique(arr, return_counts=True)
print(result)

A (array([1, 2, 3, 4]), array([2, 1, 2, 2]))

B (array([1, 2, 3, 4]), array([1, 2, 2, 2]))

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

D (array([1, 2, 3, 4]), array([1, 1, 1, 1]))

💡 Hint

Think about how many times each unique number appears in the array.

The array contains 1 twice, 2 twice, 3 twice, and 4 once. np.unique returns sorted unique elements and their counts.

❓ data_output

intermediate

Number of unique elements in a 2D array

How many unique elements are in this 2D numpy array?

NumPy

import numpy as np
arr = np.array([[1, 2, 2], [3, 1, 4], [4, 5, 5]])
unique_elements = np.unique(arr)
print(len(unique_elements))

A 6

B 4

C 7

D 5

💡 Hint

Count each distinct number in the whole 2D array.

The unique elements are 1, 2, 3, 4, 5, so total 5 unique elements. But check carefully if any number repeats or is missed.

❓ Predict Output

advanced

Output of np.unique() with return_index and return_inverse

What is the output of this code using np.unique() with return_index=true and return_inverse=true?

NumPy

import numpy as np
arr = np.array([5, 3, 5, 2, 3])
unique, index, inverse = np.unique(arr, return_index=True, return_inverse=True)
print(unique, index, inverse)

A [2 3 5] [3 1 0] [2 1 2 0 1]

B ]1 2 0 1 0[ ]0 1 3[ ]5 3 2[

C [2 3 5] [3 1 0] [0 1 0 2 1]

D [2 3 5] [3 1 0] [0 1 2 0 1]

💡 Hint

Remember that return_index gives the first occurrence index of each unique element, and return_inverse maps original elements to unique indices.

Unique sorted elements are [2, 3, 5]. Their first indices in arr are 3 for 2, 1 for 3, and 0 for 5. The inverse array maps each element in arr to the index in unique.

🔧 Debug

advanced

Identify the error in np.unique() usage

What error will this code raise?

NumPy

import numpy as np
arr = np.array([1, 2, 3])
result = np.unique(arr, return_counts='yes')
print(result)

A ValueError: invalid literal for int() with base 10: 'yes'

B TypeError: return_counts must be a boolean

C TypeError: 'str' object cannot be interpreted as an integer

D TypeError: return_counts must be a bool

💡 Hint

Check the expected type for the return_counts parameter.

The parameter return_counts expects a boolean value true or false, not a string. Passing 'yes' causes a TypeError.

🚀 Application

expert

Using np.unique() to find duplicates in a large dataset

You have a large numpy array of integers. Which code snippet correctly identifies all elements that appear more than once?

A

unique, counts = np.unique(arr, return_counts=True)
duplicates = unique[counts > 1]

B

unique = np.unique(arr)
duplicates = [x for x in unique if arr.count(x) > 1]

C duplicates = np.unique(arr[arr > 1])

D

unique, counts = np.unique(arr)
duplicates = unique[counts > 1]

💡 Hint

Use return_counts to get counts of each unique element.

Option A correctly uses np.unique with return_counts=true to get counts, then filters unique elements with counts greater than 1. Option A uses list method count which does not work on numpy arrays. Option A filters by value > 1, not duplicates. Option A tries to unpack np.unique without return_counts, causing error.