How to reorganize data in Python reshape

Method Numpy.reshape () is how the data is redefined for shape. Mister into a random array
Reshape into 5 rows and 3 columns, you can see that the remaining two columns of the first row in the (5,3) are the top two columns of the second row, and so on reshape in a column where ( -1,1) can be any column, of course, only 1, 3, 5 columns. See how the implementation is looping by rows, from the first column to the last column.
Finally, and then reshape ( -1,3), you can see that the implementation is in order to take the number of columns at a time, and then sequentially discharge.
For multidimensional arrays, the same b=np.random.random ((2,3,4,5))
C=b[:,:,:,0].reshape ( -1,1) obtained,