I have a list 'Z' with:
import numpy as np
z[0] = np.random.normal( 0, 1, ( 500, 20 ) )
z[1] = np.random.normal( 0, 1, ( 500, 30 ) )
There are about 100 arrays in the list. I am using only size 2 list for illustration. The stored arrays are all of dimension 0 of 500
I want to achieve:
C = np.concatenate( ( z[0] , z[1] ),1)
I tried:
z1 = [ np.concatenate( z[ii], 1 ) for ii in range(0,len(z)) ]
but it still returns the original list and doesn't concatenate the stored arrays
Concatenation for multidimensional arrays is somewhat ill-defined without specifying an axis along which to concatenate. I assume you want to stack your arrays horizontally because the number of rows is the same for both. The simplest call is
stacked = np.hstack(Z)
which will concatenate along axis 1. You can find documentation here.
More generally, you can also use
stacked = np.concatenate(Z, axis=1)
which works for higher-dimensional arrays, too. The corresponding documentation is here.
I was confused by the numpy stuff, but now I see what you were asking. You just have your list comprehension inside out.
Rather than
z1 = [ np.concatenate( z[ii], 1 ) for ii in range(0,len(z)) ]
You want
z1 = np.concatenate((z[ii] for ii in range(0, len(z)), 1)
Note that I changed it to a generator expresssion, as you don't really care about the intermediate list.
z rather than wrapping it in an iterator.
