np.searchsorted() for insertion points in NumPy - Time & Space Complexity

Time Complexity: np.searchsorted() for insertion points

O(m log n)

Understanding Time Complexity

We want to understand how the time needed to find insertion points grows as the input size increases when using np.searchsorted().

How does the number of operations change when we search in bigger arrays?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import numpy as np

arr = np.sort(np.random.randint(0, 1000, size=1000))
values = np.array([10, 500, 999])
indices = np.searchsorted(arr, values)
print(indices)

This code finds the positions where each value in values should be inserted in the sorted array arr to keep it sorted.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: For each value, a binary search is done on the sorted array.
• How many times: Once per value in the values array.

How Execution Grows With Input

Each search takes time that grows slowly as the array size grows, and the total time grows linearly with the number of values searched.

Input Size (n)	Approx. Operations
10	About 30 operations (3 per search x 10 values)
100	About 700 operations (7 per search x 100 values)
1000	About 10,000 operations (10 per search x 1000 values)

Pattern observation: The search per value grows slowly with array size, but total work grows roughly with the number of values searched.

Final Time Complexity

Time Complexity: O(m log n)

This means the time grows linearly with the number of values searched (m) and logarithmically with the size of the sorted array (n).

Common Mistake

[X] Wrong: "The search time grows linearly with the size of the array for each value."

[OK] Correct: The search uses binary search, which splits the array repeatedly, so it grows much slower than linear, only logarithmically.

Interview Connect

Understanding how search operations scale helps you explain efficient data lookups and insertion points, a useful skill in many data science tasks.

Self-Check

"What if the array was not sorted? How would the time complexity of finding insertion points change?"

Practice

1. What does the np.searchsorted() function do in NumPy?

easy

A. Returns the maximum value in the array

B. Sorts the array in ascending order

C. Removes duplicate values from the array

D. Finds the index where a value should be inserted to keep the array sorted

Solution

1. Step 1: Understand the purpose of np.searchsorted()

   This function finds the position where a new element can be inserted in a sorted array without breaking the order.

2. Step 2: Compare with other options

   Options B, C, and D describe different functions: sorting, removing duplicates, and finding max, which are not what searchsorted does.

3. Final Answer:

   Finds the index where a value should be inserted to keep the array sorted -> Option D

4. Quick Check:

   Insertion index finder = A [OK]

Hint: Remember: searchsorted finds insert position, not sorting [OK]

Common Mistakes:

• Confusing searchsorted with sorting functions
• Thinking it removes duplicates
• Assuming it returns values instead of indices

2. Which of the following is the correct syntax to find the insertion index of value 5 in a sorted array arr using np.searchsorted()?

easy

A. np.searchsorted(arr, value=5)

B. np.searchsorted(arr, 5)

C. arr.searchsorted(5)

D. np.searchsorted(5, arr)

Solution

1. Step 1: Recall the function signature

   The correct syntax is np.searchsorted(array, value), so the array comes first, then the value.

2. Step 2: Check each option

   np.searchsorted(arr, 5) matches the correct order. np.searchsorted(5, arr) reverses arguments. arr.searchsorted(5) is invalid because searchsorted is not a method of ndarray. np.searchsorted(arr, value=5) uses a wrong keyword argument.

3. Final Answer:

   np.searchsorted(arr, 5) -> Option B

4. Quick Check:

   Array first, value second = D [OK]

Hint: Remember: np.searchsorted(array, value) order [OK]

Common Mistakes:

• Swapping the order of arguments
• Using searchsorted as a method of array
• Using incorrect keyword arguments

3. What is the output of the following code?

import numpy as np
arr = np.array([1, 3, 5, 7])
index = np.searchsorted(arr, 4)
print(index)

medium

A. 2

B. 1

C. 3

D. 4

Solution

1. Step 1: Understand the array and value

   The array is [1, 3, 5, 7], and we want to insert 4 while keeping it sorted.

2. Step 2: Find the insertion index

   4 fits between 3 (index 1) and 5 (index 2), so the insertion index is 2.

3. Final Answer:

   2 -> Option A

4. Quick Check:

   Insert 4 between 3 and 5 = 2 [OK]

Hint: Find where value fits between sorted elements [OK]

Common Mistakes:

• Choosing index of smaller element
• Choosing index of larger element
• Confusing zero-based indexing

4. The code below throws an error. What is the mistake?

import numpy as np
arr = np.array([2, 4, 6, 8])
index = np.searchsorted(arr, side='left', 5)
print(index)

medium

A. The 'side' argument should come after the value argument

B. The array must be sorted in descending order

C. The function does not accept keyword arguments

D. The value to insert must be an array, not a scalar

Solution

1. Step 1: Check function argument order

   np.searchsorted expects the array first, then the value, then optional keywords like side.

2. Step 2: Identify the error in argument placement

   The code passes side='left' before the value 5, which is incorrect syntax.

3. Final Answer:

   The 'side' argument should come after the value argument -> Option A

4. Quick Check:

   Keyword args after positional args = A [OK]

Hint: Put keyword arguments after positional ones [OK]

Common Mistakes:

• Placing keyword arguments before positional arguments
• Assuming array must be descending
• Thinking scalar values are invalid

5. Given a sorted array arr = np.array([1, 2, 2, 3, 4]), which code snippet will insert the value 2 after all existing 2s using np.searchsorted()?

hard

A. index = np.searchsorted(arr, 2)

B. index = np.searchsorted(arr, 2, side='left')

C. index = np.searchsorted(arr, 2, side='right')

D. index = np.searchsorted(arr, 2, side='both')

Solution

1. Step 1: Understand the side parameter

   side='right' returns the insertion index after existing equal values; side='left' inserts before.

2. Step 2: Apply to the array

   For value 2 in [1, 2, 2, 3, 4], side='right' gives index 3, after the two 2s.

3. Final Answer:

   index = np.searchsorted(arr, 2, side='right') -> Option C

4. Quick Check:

   Insert after equals = side='right' = C [OK]

Hint: Use side='right' to insert after equal values [OK]

Common Mistakes:

• Using side='left' inserts before equal values
• Assuming default side inserts after equals
• Using invalid side='both'