We can select elements and subarrays from NumPy arrays using the standard square bracket notation that we use with Python lists.

One dimensional arrays

Along a single axis, we use integers to select single elements.

We can use -ve integers to index elements from the end of an array. The last element is indexed with -1, the second last element is indexed with -2 and so on.

import numpy as np

data = np.arange(0, 10, 2)

data
array([0, 2, 4, 6, 8])

data[1]  # the second element, at index 1
# 2

data[-1] # the last element
# 8

data[2]
# 4

We can select a sequence of elements using slices, which are specified using the colon (:) notation.

Array slicing expressions

• a[x:y] - Select elements with index starting at x and ending at y-1. (x & y are integers.). The slice a[x:y] can also be written as a[x:y:1], where the number 1 specifies that every element between x & y should be selected. To select every second element between x & y use a[x:y:2].

• a[:] - Select all the elements in the given axis.

• a[:x ] - Select elements starting with 0 and going up to index x-1 (integer).

• a[x:] - Select elements starting from index x (integer) and going up to the last element in the array.

• a[::-1] - Select all elements in the reverse order.

import numpy as np

data = np.arange(0, 10, 2)

data
# array([0, 2, 4, 6, 8])

data[1:3]
# array([2, 4])

data[:4]
# array([0, 2, 4, 6])

data[:]
# array([0, 2, 4, 6, 8])

Multidimensional arrays

We can extract columns and rows from a two-dimensional array using a combination of slices and integer indexing.

import numpy as np

data = np.array([[1, 4, 7, 8, 3], [10, 45, 24, 39, 87], [23, 44, 78, 19, 39], [74, 38, 26, 38, 45]])

data
# array([[ 1,  4,  7,  8,  3],
#        [10, 45, 24, 39, 87],
#        [23, 44, 78, 19, 39],
#        [74, 38, 26, 38, 45]])

data[:, 2] # select the third column
# array([ 7, 24, 78, 26])

data[2, :] # select the third row
# array([23, 44, 78, 19, 39])

We can extract subarrays by applying a slice on each of the array dimensions.

data[1:, :2]
# array([[10, 45],
#        [23, 44],
#        [74, 38]])