numpy array concatenate: "ValueError: all the input arrays must have same number of dimensions"

Question

How to concatenate these numpy arrays?

first np.array with a shape (5,4)

[[  6487    400 489580      0]
 [  6488    401 492994      0]
 [  6491    408 489247      0]
 [  6491    408 489247      0]
 [  6492    402 499013      0]]

second np.array with a shape (5,)

[  16.   15.   12.  12.  17. ]

final result should be

[[  6487    400    489580    0   16]
 [  6488    401    492994    0   15]
 [  6491    408    489247    0   12]
 [  6491    408    489247    0   12]
 [  6492    402    499013    0   17]]

I tried np.concatenate([array1, array2]) but i get this error

ValueError: all the input arrays must have same number of dimensions

What am I doing wrong?

Answer 1

To use np.concatenate , we need to extend the second array to 2D and then concatenate along axis=1 -

np.concatenate((a,b[:,None]),axis=1)

Alternatively, we can use np.column_stack that takes care of it -

np.column_stack((a,b))

Sample run -

In [84]: a
Out[84]: 
array([[54, 30, 55, 12],
       [64, 94, 50, 72],
       [67, 31, 56, 43],
       [26, 58, 35, 14],
       [97, 76, 84, 52]])

In [85]: b
Out[85]: array([56, 70, 43, 19, 16])

In [86]: np.concatenate((a,b[:,None]),axis=1)
Out[86]: 
array([[54, 30, 55, 12, 56],
       [64, 94, 50, 72, 70],
       [67, 31, 56, 43, 43],
       [26, 58, 35, 14, 19],
       [97, 76, 84, 52, 16]])

If b is such that its a 1D array of dtype=object with a shape of (1,) , most probably all of the data is contained in the only element in it, we need to flatten it out before concatenating. For that purpose, we can use np.concatenate on it too. Here's a sample run to make the point clear -

In [118]: a
Out[118]: 
array([[54, 30, 55, 12],
       [64, 94, 50, 72],
       [67, 31, 56, 43],
       [26, 58, 35, 14],
       [97, 76, 84, 52]])

In [119]: b
Out[119]: array([array([30, 41, 76, 13, 69])], dtype=object)

In [120]: b.shape
Out[120]: (1,)

In [121]: np.concatenate((a,np.concatenate(b)[:,None]),axis=1)
Out[121]: 
array([[54, 30, 55, 12, 30],
       [64, 94, 50, 72, 41],
       [67, 31, 56, 43, 76],
       [26, 58, 35, 14, 13],
       [97, 76, 84, 52, 69]])

Answer 2

There's also np.c_

>>> a = np.arange(20).reshape(5, 4)
>>> b = np.arange(-1, -6, -1)
>>> a
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11],
       [12, 13, 14, 15],
       [16, 17, 18, 19]])                                                                                                                                   
>>> b                                                                                                                                                       
array([-1, -2, -3, -4, -5])                                                                                                                                 
>>> np.c_[a, b]
array([[ 0,  1,  2,  3, -1],          
       [ 4,  5,  6,  7, -2],                       
       [ 8,  9, 10, 11, -3],                      
       [12, 13, 14, 15, -4],                                
       [16, 17, 18, 19, -5]])

Answer 3

You can do something like this.

import numpy as np

x = np.random.randint(100, size=(5, 4))
y = [16, 15, 12, 12, 17]

print(x)

val = np.concatenate((x,np.reshape(y,(x.shape[0],1))),axis=1)
print(val)

This outputs:

[[32 37 35 53]
 [64 23 95 76]
 [17 76 11 30]
 [35 42  6 80]
 [61 88  7 56]]

[[32 37 35 53 16]
 [64 23 95 76 15]
 [17 76 11 30 12]
 [35 42  6 80 12]
 [61 88  7 56 17]]

Answer 4

It is not related to the original question. But it can be useful when adding columns vertically in a loop.

a = np.empty([0,1], dtype=float) 
b = np.array([[6487, 400, 489.580],
              [6488, 401, 492.994],
              [6491, 408, 489.247],
              [6491, 408, 489.247],
              [6492, 402, 499.013]])

for c in range(0, 3):
    col = b[:, c]
    col = col.reshape(len(col),1)
    a = np.vstack((a, col))

print(a)

----------------------------------------------
Output:
[[6487.   ]
 [6488.   ]
 [6491.   ]
 [6491.   ]
 [6492.   ]
 [ 400.   ]
 [ 401.   ]
 [ 408.   ]
 [ 408.   ]
 [ 402.   ]
 [ 489.58 ]
 [ 492.994]
 [ 489.247]
 [ 489.247]
 [ 499.013]]