6 Numpy (Numeric Python)

This is a package which provides some types and functions for maths. The key use of numpy is the numpy.array() function, which provides something similar to the c() function (atomic vector) in R. Unlike lists (in both R and Python), the numpy.array() function allows vectorised operations.

By convention, numpy is normally imported as np.

import numpy as np

6.1 Numpy arrays

To create a numpy array:

import numpy as np
temp = np.array([1,2,3,4])
print(temp)

## [1 2 3 4]

Note the strange bracket nesting - it looks like the array function takes a list as an argument, which means you need to use the square brackets to make a list, then pass this list as the argument to the array function.

One dimensional numpy arrays are useful for R users, because they behave like atomic vectors in R; in fact numpy arrays are also atomic - they can only hold items with the same data type.

6.2 Two dimensional numpy arrays

You can define n-dimensional numpy arrays using lists of lists. The highest level of the hierarchy is the first dimension of the n-dimensional array, the second level is the second dimension of the array, and so on. When working with a two dimensional array, this means that you can create a 2 row, 5 column numpy array using the following command:

import numpy as np
np_2d = np.array([[1, 2, 3, 4, 5 ],
[6, 7, 8, 9, 10]])
print(np_2d)

## [[ 1  2  3  4  5]
##  [ 6  7  8  9 10]]

To interrogate the properties of the array you can use methods. For example, you can get the dimensions of the array using np_2d.shape - this would be written as dim(np_2d) in R.

print(np_2d.shape)

## (2, 5)