Concatenate NumPy 2D array with column (1D array)

Question

Suppose I have a 2D NumPy array values. I want to add new column to it. New column should be values[:, 19] but lagged by one sample (first element equals to zero). It could be returned as np.append([0], values[0:-2:1, 19]). I tried: Numpy concatenate 2D arrays with 1D array

temp = np.append([0], [values[1:-2:1, 19]])
values = np.append(dataset.values, temp[:, None], axis=1)

but I get:

ValueError: all the input array dimensions except for the concatenation axis 
must match exactly

I tried using c_ too as:

temp = np.append([0], [values[1:-2:1, 19]])
values = np.c_[values, temp]

but effect is the same. How this concatenation could be made. I think problem is in temp orientation - it is treated as a row instead of column, so there is an issue with dimensions. In Octave ' (transpose operator) would do the trick. Maybe there is similiar solution in NumPy?

Anyway, thank you for you time.

Best regards,
Max

If `temp` is shape (n,) (1d), you can make a 2d 'column' with `temp[:,None]` or `temp.reshape(n,1)`. That can then be `concatenate` with a 2d (n,m) array, on axis=1. The key difference with Octave is that numpy arrays can be true 1d. — hpaulj, Aug 19 '19 at 21:27

score 2 · Accepted Answer · answered Aug 20 '19 at 00:37

In [76]: values = np.arange(16).reshape(4,4)                                                                 
In [77]: temp = np.concatenate(([0], values[1:,-1]))                                                         
In [78]: values                                                                                              
Out[78]: 
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11],
       [12, 13, 14, 15]])
In [79]: temp                                                                                                
Out[79]: array([ 0,  7, 11, 15])

This use of concatenate to make temp is similar to your use of append (which actually uses concatenate).

Sounds like you want to join values and temp in this way:

In [80]: np.concatenate((values, temp[:,None]),axis=1)                                                       
Out[80]: 
array([[ 0,  1,  2,  3,  0],
       [ 4,  5,  6,  7,  7],
       [ 8,  9, 10, 11, 11],
       [12, 13, 14, 15, 15]])

Again I prefer using concatenate directly.

score 0 · Answer 2 · answered Aug 19 '19 at 22:19

You need to convert the 1D array to 2D as shown. You can then use vstack or hstack with reshaping to get the final array you want as shown:

a = np.array([[1, 2, 3],[4, 5, 6]])
b = np.array([[7, 8, 9]])
c = np.vstack([ele for ele in [a, b]])
print(c)
c = np.hstack([a.reshape(1,-1) for a in [a,b]]).reshape(-1,3)
print(c)

Either way, the output is:

[[1 2 3] [4 5 6] [7 8 9]]

Hope I understood the question correctly

Concatenate NumPy 2D array with column (1D array)

2 Answers2