np.save() and np.load() for binary in NumPy - Time & Space Complexity

Time Complexity: np.save() and np.load() for binary

O(n)

Understanding Time Complexity

We want to understand how the time to save and load data with numpy changes as the data size grows.

How does the time cost grow when saving or loading bigger arrays?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import numpy as np

n = 1000  # example size
arr = np.arange(n)  # create an array of size n
np.save('data.npy', arr)  # save array to binary file
loaded_arr = np.load('data.npy')  # load array from binary file

This code creates an array of size n, saves it to a binary file, then loads it back into memory.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Reading and writing each element of the array to disk.
• How many times: Once for each element in the array, so n times.

How Execution Grows With Input

As the array size grows, the time to save or load grows roughly in direct proportion.

Input Size (n)	Approx. Operations
10	10 operations (reading/writing 10 elements)
100	100 operations
1000	1000 operations

Pattern observation: Doubling the input size roughly doubles the time needed.

Final Time Complexity

Time Complexity: O(n)

This means the time to save or load grows linearly with the number of elements.

Common Mistake

[X] Wrong: "Saving or loading is instant regardless of data size."

[OK] Correct: The process must handle each element, so bigger arrays take more time.

Interview Connect

Knowing how data saving and loading scales helps you understand performance in real projects.

Self-Check

"What if we compressed the file while saving? How would the time complexity change?"

Practice

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

easy

A. Saves a NumPy array to a binary file on disk

B. Loads a NumPy array from a binary file

C. Converts a NumPy array to a list

D. Prints the contents of a NumPy array

Solution

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

   The np.save() function is designed to save a NumPy array to a file in binary format, preserving its data type and shape.

2. Step 2: Differentiate from np.load()

   np.load() is used to load arrays from files, not save them. Other options do not relate to saving files.

3. Final Answer:

   Saves a NumPy array to a binary file on disk -> Option A

4. Quick Check:

   np.save() saves array [OK]

Hint: np.save() writes array to file, np.load() reads it back [OK]

Common Mistakes:

• Confusing np.save() with np.load()
• Thinking np.save() converts array to list
• Assuming np.save() prints array

2. Which of the following is the correct syntax to save a NumPy array arr to a file named data.npy?

easy

A. np.savefile('data.npy', arr)

B. np.save(arr, 'data.npy')

C. np.load('data.npy', arr)

D. np.save('data.npy', arr)

Solution

1. Step 1: Recall np.save() parameter order

   The first argument is the filename (string), the second is the array to save.

2. Step 2: Check other options for correctness

   np.save(arr, 'data.npy') reverses parameters, np.load('data.npy', arr) uses np.load() which loads, not saves, np.savefile('data.npy', arr) uses a non-existent function.

3. Final Answer:

   np.save('data.npy', arr) -> Option D

4. Quick Check:

   Filename first, array second in np.save() [OK]

Hint: np.save(filename, array) always filename first [OK]

Common Mistakes:

• Swapping filename and array arguments
• Using np.load() instead of np.save() to save
• Using wrong function name like np.savefile()

3. What will be the output of the following code?

import numpy as np
arr = np.array([1, 2, 3])
np.save('file.npy', arr)
loaded_arr = np.load('file.npy')
print(loaded_arr)

medium

A. [1 2 3]

B. ['1' '2' '3']

C. Error: file not found

D. [[1 2 3]]

Solution

1. Step 1: Save and load the array

   The array [1, 2, 3] is saved to 'file.npy' and then loaded back exactly as it was.

2. Step 2: Understand print output of loaded array

   Printing the loaded array shows the original array as [1 2 3] without quotes or extra brackets.

3. Final Answer:

   [1 2 3] -> Option A

4. Quick Check:

   np.load(np.save()) returns original array [OK]

Hint: np.load(np.save()) returns original array unchanged [OK]

Common Mistakes:

• Expecting string elements instead of integers
• Thinking np.load() returns nested arrays
• Assuming file not found error without saving first

4. What is wrong with this code snippet?

import numpy as np
arr = np.array([4, 5, 6])
np.save('mydata.npy')
loaded = np.load('mydata.npy')
print(loaded)

medium

A. np.save() cannot save integer arrays

B. np.load() should be called before np.save()

C. np.save() is missing the array argument to save

D. The filename should have .txt extension

Solution

1. Step 1: Check np.save() usage

   The np.save() function requires two arguments: filename and array. Here, the array argument is missing.

2. Step 2: Verify other options

   np.load() should be called before np.save() is incorrect because loading happens after saving. The filename should have .txt extension is wrong because .npy is the correct extension. np.save() cannot save integer arrays is false; np.save() can save integer arrays.

3. Final Answer:

   np.save() is missing the array argument to save -> Option C

4. Quick Check:

   np.save() needs filename and array [OK]

Hint: np.save() always needs array argument after filename [OK]

Common Mistakes:

• Forgetting to pass the array to np.save()
• Thinking .txt is needed instead of .npy
• Confusing order of np.save() and np.load()

5. You have saved multiple arrays separately using np.save() as arr1.npy and arr2.npy. How can you load both arrays and combine them into a single 2D array where each original array is a row?

hard

A. Use np.load('arr1.npy', 'arr2.npy') directly

B. Load each with np.load() and use np.vstack([arr1, arr2])

C. Save both arrays in one file using np.save() and then load

D. Load arrays and use np.concatenate(arr1, arr2, axis=1)

Solution

1. Step 1: Load arrays separately

   Since arrays are saved in separate files, load each using np.load() individually.

2. Step 2: Combine arrays as rows

   Use np.vstack([arr1, arr2]) to stack arrays vertically, making each array a row in the new 2D array.

3. Step 3: Check other options

   Use np.load('arr1.npy', 'arr2.npy') directly is invalid syntax, Save both arrays in one file using np.save() and then load is incorrect because np.save() saves one array per file, Load arrays and use np.concatenate(arr1, arr2, axis=1) concatenates along columns which may not work if shapes differ.

4. Final Answer:

   Load each with np.load() and use np.vstack([arr1, arr2]) -> Option B

5. Quick Check:

   Load separately, stack with vstack [OK]

Hint: Load arrays separately, stack rows with np.vstack() [OK]

Common Mistakes:

• Trying to load multiple files in one np.load() call
• Using np.concatenate with wrong axis
• Assuming np.save() can save multiple arrays in one file