python: join two 2 dimensional numpy array to create 3 dimentional array

Question

I want to "join" two 2-dimensional numpy arrays of same shape to create a dimensional numpy array. I can easily do this using loop but I am looking for faster way. Here is the toy example. The two numpy arrays are:

data1 = np.array([[1,2,3,4,5],[6,7,8,9,10],[11,12,13,14,15]])
data2 = np.array([[1,2,3,4,5],[6,7,8,9,10],[11,12,13,14,15]])*100

data1

> array([[ 1,  2,  3,  4,  5],
         [ 6,  7,  8,  9, 10],
         [11, 12, 13, 14, 15]])

data2

> array([[ 100,  200,  300,  400,  500],
         [ 600,  700,  800,  900, 1000],
         [1100, 1200, 1300, 1400, 1500]])

Both have a shape (3,5). I want to create (3,5,2) shape numpy array. That is:

data3 = []
for irow in range(data1.shape[0]):
    data3_temp = []
    for icol in range(data1.shape[1]):
        data3_temp.append([data1[irow,icol],
                           data2[irow,icol]])
    data3.append(data3_temp)
data3 = np.array(data3)

data3.shape
> (3, 5, 2)

data3

>array([[[   1,  100],
    [   2,  200],
    [   3,  300],
    [   4,  400],
    [   5,  500]],

   [[   6,  600],
    [   7,  700],
    [   8,  800],
    [   9,  900],
    [  10, 1000]],

   [[  11, 1100],
    [  12, 1200],
    [  13, 1300],
    [  14, 1400],
    [  15, 1500]]])

Please let me know.

Serenity · Accepted Answer · 2017-03-25 04:12:36Z

3

Use numpy.dstack to stack arrays in sequence depth wise (along third axis): https://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.dstack.html

import numpy as np

data1 = np.array([[1,2,3,4,5],[6,7,8,9,10],[11,12,13,14,15]])
data2 = np.array([[1,2,3,4,5],[6,7,8,9,10],[11,12,13,14,15]])*100

print np.dstack((data1,data2))

Output:

[[[   1  100]
  [   2  200]
  [   3  300]
  [   4  400]
  [   5  500]]

 [[   6  600]
  [   7  700]
  [   8  800]
  [   9  900]
  [  10 1000]]

 [[  11 1100]
  [  12 1200]
  [  13 1300]
  [  14 1400]
  [  15 1500]]]

answered Mar 25, 2017 at 4:12

Serenity

37.1k21 gold badges125 silver badges117 bronze badges

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

2 Comments

hpaulj Over a year ago

And now with a bit more flexibility, np.stack((), axis=-1).

kmario23 Over a year ago

@hpaulj nice trick! Particularly for this question, np.stack((), axis=2) is much clear. But, using axis=-1 is applicable in a general setting :)

plasmon360 · Accepted Answer · 2017-03-25 04:20:06Z

1

one method:

data3= np.zeros([ data1.shape[0],data1.shape[1], 2])

data3[:,:,0],data3[:,:,1] = data1,data2

print data3.shape
print data3

results in

(3L, 5L, 2L)

>array([[[   1,  100],
    [   2,  200],
    [   3,  300],
    [   4,  400],
    [   5,  500]],

   [[   6,  600],
    [   7,  700],
    [   8,  800],
    [   9,  900],
    [  10, 1000]],

   [[  11, 1100],
    [  12, 1200],
    [  13, 1300],
    [  14, 1400],
    [  15, 1500]]])

answered Mar 25, 2017 at 4:20

plasmon360

4,1991 gold badge21 silver badges19 bronze badges

Comments

Paul Panzer · Accepted Answer · 2017-03-25 05:29:16Z

1

Or you can use the r_ concatenator:

np.r_['2,3,0', data1, data2]
# array([[[   1,  100],
          [   2,  200],
          [   3,  300],
          [   4,  400],
          [   5,  500]],

         [[   6,  600],
          [   7,  700],
          [   8,  800],
          [   9,  900],
          [  10, 1000]],

         [[  11, 1100],
          [  12, 1200],
          [  13, 1300],
          [  14, 1400],
          [  15, 1500]]])

answered Mar 25, 2017 at 5:29

Paul Panzer

53.3k3 gold badges60 silver badges103 bronze badges

Collectives™ on Stack Overflow

python: join two 2 dimensional numpy array to create 3 dimentional array

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related