Here are the most common ways to save a NumPy array to a file, depending on what you plan to do with the data later.
（图片来源网络，侵删）
Save as a .npy file (Best for Python/NumPy)
This is the standard way to save a single NumPy array. It preserves the data type and shape perfectly.
import numpy as np
arr = np.array([1, 2, 3, 4, 5])
# Save the array
np.save('my_array.npy', arr)
# Load the array back
loaded_arr = np.load('my_array.npy')
print(loaded_arr)
Save as a .npz file (Best for multiple arrays)
If you have more than one array, you can compress them into a single .npz file.
import numpy as np
arr1 = np.array([1, 2, 3])
arr2 = np.array([4, 5, 6])
# Save multiple arrays
np.savez('multiple_arrays.npz', array1=arr1, array2=arr2)
# Load them back
data = np.load('multiple_arrays.npz')
print(data['array1'])
print(data['array2'])
Save as a .csv or .txt file (Best for Excel or other languages)
If you need to open the array in Excel, MATLAB, or read it with a language other than Python, save it as a text file.
import numpy as np
arr = np.array([[1, 2, 3], [4, 5, 6]])
# Save as a CSV file
np.savetxt('my_array.csv', arr, delimiter=',')
# Load it back
loaded_arr = np.loadtxt('my_array.csv', delimiter=',')
print(loaded_arr)
Save as a .csv using Pandas (Best for data with headers)
If your array represents tabular data and you need column or row headers, use the Pandas library.
（图片来源网络，侵删）
import pandas as pd
import numpy as np
arr = np.array([[1, 2, 3], [4, 5, 6]])
# Save with column names
df = pd.DataFrame(arr, columns=['A', 'B', 'C'])
df.to_csv('my_data.csv', index=False)
# Load it back
loaded_df = pd.read_csv('my_data.csv')
print(loaded_df.values) # .values converts it back to a NumPy array
Save as a .json file (Best for web APIs and JavaScript)
If you need to pass the array to a web frontend or a NoSQL database, JSON is the best format.
import json
import numpy as np
arr = np.array([1, 2, 3, 4, 5])
# Save as JSON
with open('my_array.json', 'w') as f:
json.dump(arr.tolist(), f) # .tolist() converts the NumPy array to a standard Python list
# Load it back
with open('my_array.json', 'r') as f:
loaded_list = json.load(f)
loaded_arr = np.array(loaded_list)
print(loaded_arr)
Summary: Which should you use?
- Loading back into Python? Use
.npy(Fastest and smallest file size). - Multiple arrays? Use
.npz. - Opening in Excel? Use
.csv(np.savetxtorpandas). - Sending to a web browser? Use
.json.
