Time for action – slicing and indexing multidimensional arrays

The ndarray class supports slicing over multiple dimensions. For convenience, we refer to many dimensions at once, with an ellipsis.

1. To illustrate, create an array with the arange() function and reshape it:

   In: b = arange(24).reshape(2,3,4)
   In: b.shape
   Out: (2, 3, 4)
   In: b
   Out:
   array([[[ 0,  1,  2,  3],
           [ 4,  5,  6,  7],
           [ 8,  9, 10, 11]],
          [[12, 13, 14, 15],
           [16, 17, 18, 19],
           [20, 21, 22, 23]]])
   

   The array b has 24 elements with values 0 to 23 and we reshaped it to be a two-by-three-by-four, three-dimensional array. We can visualize this as a two-story building with 12 rooms on each floor, 3 rows and 4 columns (alternatively we can think of it as a spreadsheet with sheets, rows, and columns). As you have probably guessed, the reshape() function changes the shape of an array. We give it a tuple of integers, corresponding to the new shape. If the dimensions are not compatible with the...