CODEWITHSUNDEEP
pythonnumpysave
Cicilio
Cicilio

Reputation: 432

How to append an array to an existing `.npz` file?

I have the following code, which generate the mat file (in .npz format):

import numpy as np
x = np.arange(10)
np.savez('mat',x)

Now I want to append another array to the existing file:

y = np.arange(21)
np.savez('mat',y)

But, it replaces the x array with y. I want to have both of these arrays in the file. Besides, I do not want to write them to the file at the same time.

Upvotes: 7

Views: 5736

Answers (3)

Alexandar Zeng
Alexandar Zeng

Reputation: 13

Not a real solution here but I'd like to share a way to add array to a new nyz file.

import numpy as np

array_dict = {'new_array': new_array}
for k in old_npz_file:
    array_dict[k] = old_npz_file[k]

np.savez(file_path, **array_dict)

I think you could say it's one way to "add" array.

Upvotes: 1

Hamzah Al-Qadasi
Hamzah Al-Qadasi

Reputation: 9786

Step1: create your first file

d = np.arange(10)
np.savez("mat",x = d)  # do not forget to name your variable d

Step2: read your file, convert it to a dictionary object, add your new variable to the dictionary, and finally overwrite your mat.npz with this dictionary

data = np.load("mat.npz")
data = dict(data)
data["y"] = np.arange(21)
np.savez("mat",**data)

Step 3: to verify if the last variable is added to the mat.npz file.

data = np.load("mat.npz")
print(data[x])   # [0 1 2 3 4 5 6 7 8 9]
print(data[y])   # [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20]

Upvotes: 3

Tonsic
Tonsic

Reputation: 1138

I took the comments on the post as a start and the following code seems to do what you want. (I'm using python 3.6 and base modules only, it probably works for other python versions)

import zipfile
import io
import os
import numpy as np

test_np1 = np.random.random(100)
test_np2 = np.random.random(100)

# initial file
filename = 'npfile.npz'
try:
    np.savez_compressed(filename, test_np1=test_np1)

    with np.load(filename) as loaded_npz:
        print(np.array_equal(test_np1, loaded_npz['test_np1']))

    bio = io.BytesIO()
    np.save(bio, test_np2)
    with zipfile.ZipFile(filename, 'a') as zipf:
        # careful, the file below must be .npy
        zipf.writestr('test_np2.npy', data=bio.getbuffer().tobytes())

    with np.load(filename) as loaded_npz:
        print(np.array_equal(test_np1, loaded_npz['test_np1']))
        print(np.array_equal(test_np2, loaded_npz['test_np2']))
finally:
    os.remove(filename)

Upvotes: 3

Related Questions