Partial sorting with np.partition() in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of np.partition() on a 1D array

What is the output of the following code using np.partition() on a 1D array?

NumPy

import numpy as np
arr = np.array([7, 2, 5, 3, 9, 1])
result = np.partition(arr, 3)
print(result)

A [1 2 3 5 7 9]

B [2 1 3 5 7 9]

C [1 2 3 7 5 9]

D [2 1 3 5 9 7]

💡 Hint

Remember, np.partition rearranges so that the element at the kth position is in sorted order, and all smaller elements are before it, but the rest are unordered.

The np.partition(arr, 3) places the element that would be at index 3 in a sorted array at index 3. Elements before index 3 are smaller or equal, but not necessarily sorted. Elements after index 3 are larger or equal, unordered.

Sorted array is [1, 2, 3, 5, 7, 9]. The element at index 3 is 5. So after partition, 5 is at index 3, smaller elements (1,2,3) are before it in any order, and larger elements (7,9) after it in any order. Option D matches this.

❓ data_output

intermediate

Number of elements less than pivot after np.partition()

After running np.partition(arr, 2) on arr = np.array([10, 4, 5, 8, 6, 11, 3]), how many elements are guaranteed to be less than or equal to the element at index 2 in the result?

NumPy

import numpy as np
arr = np.array([10, 4, 5, 8, 6, 11, 3])
result = np.partition(arr, 2)
print(result)

A At least 2 elements

B Exactly 3 elements

C At least 3 elements

D Exactly 2 elements

💡 Hint

The kth element after partition is in its sorted position, so all elements before it are smaller or equal.

Using np.partition(arr, 2) places the element that would be at index 2 in sorted order at index 2. All elements before index 2 are smaller or equal to it. So there are at least 3 elements (indices 0,1,2) less than or equal to that element.

🔧 Debug

advanced

Identify the error in np.partition usage

What error does the following code raise?

NumPy

import numpy as np
arr = np.array([1, 2, 3, 4, 5])
result = np.partition(arr, 10)
print(result)

A ValueError: kth out of bounds

B TypeError: invalid type for kth

C IndexError: index 10 is out of bounds for axis 0 with size 5

D No error, returns the sorted array

💡 Hint

Check if the kth index is valid for the array size.

The kth argument must be within the array bounds. Here, 10 is larger than the array length 5, so np.partition raises a ValueError.

🚀 Application

advanced

Using np.partition() to find the 3 smallest values

Which option correctly extracts the 3 smallest values from arr = np.array([12, 3, 5, 7, 19, 1]) using np.partition()?

NumPy

import numpy as np
arr = np.array([12, 3, 5, 7, 19, 1])

A np.partition(arr, 2)[:3]

B np.partition(arr, 3)[:3]

C np.partition(arr, 3)[3:]

D np.partition(arr, 2)[3:]

💡 Hint

Remember, kth index is zero-based, so to get 3 smallest elements, use k=2.

To get the 3 smallest elements, partition at index 2 (third element). The first 3 elements after partition contain the 3 smallest values, unordered.

🧠 Conceptual

expert

Behavior of np.partition() on 2D arrays

Given a 2D array arr = np.array([[9, 4, 7], [1, 3, 5]]), what does np.partition(arr, 1, axis=1) do?

NumPy

import numpy as np
arr = np.array([[9, 4, 7], [1, 3, 5]])
result = np.partition(arr, 1, axis=1)
print(result)

A Partitions each column so the element at index 1 is in sorted position within that column

B Partitions each row so the element at index 1 is in sorted position within that row

C Sorts the entire array along axis 1

D Raises an error because axis=1 is invalid for partition

💡 Hint

Check the meaning of axis=1 for a 2D array in numpy.

Axis=1 means partition is done along rows. Each row is partitioned so that the element at index 1 is in its sorted position within that row, with smaller elements before it and larger after it.