NumPy Saving/Loading with .npy Files

Saving NumPy Arrays with np.save()

The np.save() function saves a NumPy array to a file in .npy format — a binary format that's efficient and preserves all array metadata (shape, dtype, etc.). This is ideal for caching intermediate results or storing data for later use.

Basic Usage

import numpy as np

arr = np.array([[10, 20, 30],
                [40, 50, 60]])

# Save array to 'data.npy' in current directory
np.save('data.npy', arr)

print("Array saved to 'data.npy'")

import numpy as np

arr = np.array([[10, 20, 30],
                [40, 50, 60]])

# Save array to 'data.npy' in current directory
np.save('data.npy', arr)

print("Array saved to 'data.npy'")

What This Does

• 'data.npy' is the filename where the array is saved.
• The file is saved in binary format (not human-readable).
• No file extension is automatically added if one is present, but .npy is appended if omitted.

Output

Array saved to 'data.npy'

Array saved to 'data.npy'

💡 Tip: If you're saving arrays to a specific path, make sure the folder exists or handle it with os.makedirs().

Loading NumPy Arrays with np.load()

To read a NumPy array back into memory, use the np.load() function. This works with any array saved using np.save() and reconstructs the original array, including its shape and data type.

Basic Usage

import numpy as np

# Load array from file
loaded_arr = np.load('data.npy')

print("Loaded array:")
print(loaded_arr)

import numpy as np

# Load array from file
loaded_arr = np.load('data.npy')

print("Loaded array:")
print(loaded_arr)

How It Works

• Reconstructs the array from the binary .npy file.
• Works even for multi-dimensional or complex-typed arrays.
• Returns a fully functional ndarray object.

Output

Loaded array:
[[10 20 30]
 [40 50 60]]

Loaded array:
[[10 20 30]
 [40 50 60]]

💡 Tip: Make sure the file path is correct. If the file doesn't exist or is corrupted, np.load() will raise an error.

Frequently Asked Questions