Append to the numpy array another numpy array as array, not it's elements

Question

I want to to make an array of arrays, but when I use np.append I get the list if their elements:

import numpy as np
im_data = np.array(['image0','image1','image2','image3','image4','image5','image6','image7','image8','image9','image10','image11','image12','image13','image14'])
batches = [[1,2,3,4],[7,8,9,10],[3,4,5,6]]

image_batches = []
for batch in batches:
    image_batches = np.append(image_batches,[im_data[batch]])

This is what I get:

In: image_batches

Out: array(['image1', 'image2', 'image3', 'image4', 'image7', 'image8', 'image9', 'image10', 'image3', 'image4', 'image5', 'image6'], dtype='<U32')

and this is what I need:

array([['image1', 'image2', 'image3', 'image4'], ['image7', 'image8', 'image9', 'image10'], ['image3', 'image4', 'image5', 'image6']], dtype='<U7')

I achieved this by using

image_batches = im_data[batches[0]]
for batch in batches[1:]:
    image_batches = np.vstack([image_batches, im_data[batch]])

but maybe there is more elegant way to do it?

Can't you make one list with all those im_data[batch]. Repeated np.append or vstack in a loop is inefficient. Do one vstack on a list. You already know how to make an array from a list of items (or list of lists). — hpaulj
– hpaulj, Commented Jul 27, 2020 at 23:11
Does im_data[batches] do anything useful? If you can index with a sublist, I think you can do the same with the whole list. — hpaulj
– hpaulj, Commented Jul 28, 2020 at 0:21

Ehsan · Accepted Answer · 2020-07-28 01:18:34Z

2

Like mentioned by @hpaulj in comments, you can directly use advances indexing:

im_data[np.array(batches)]

output:

[['image1' 'image2' 'image3' 'image4']
 ['image7' 'image8' 'image9' 'image10']
 ['image3' 'image4' 'image5' 'image6']]

answered Jul 28, 2020 at 1:18

Ehsan

12.5k2 gold badges24 silver badges36 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Michael · Accepted Answer · 2020-07-28 10:22:27Z

You can use list comprehension and then convert it into numpy.array:

import numpy as np
im_data = np.array(['image0','image1','image2','image3','image4','image5','image6','image7','image8','image9','image10','image11','image12','image13','image14'])
batches = [[1,2,3,4],[7,8,9,10],[3,4,5,6]]

image_batches = np.array([im_data[batch_idx] for batch_idx in batches])
image_batches

Output:

array([['image1', 'image2', 'image3', 'image4'],
       ['image7', 'image8', 'image9', 'image10'],
       ['image3', 'image4', 'image5', 'image6']], dtype='<U7')

Cheers.

Collectives™ on Stack Overflow

Append to the numpy array another numpy array as array, not it's elements

2 Answers 2

Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Related