NumPy Random Generator | permutation method
Start your free 7-days trial now!
NumPy Random Generator's permutation(~) method return a new array with the values shuffled.
To shuffle values in-place, use shuffle(~).
Also, the difference between permutation(~) and permuted(~) is that the former shuffles rows or columns for two-dimensional arrays, but permuted(~) shuffles values independent of the other rows or columns. Consult examples below for clarification.
Parameters
1. x | int or array-like
If
xis anint, then anp.arange(x)is randomly shuffled and returned.If
xisarray-like, then a new array with randomly shuffled values is returned.
2. axis | int | optional
The axis by which to perform the shuffling. By default, axis=0.
Return Value
A NumPy array.
Examples
Passing an integer
To get a shuffled array of [0,1,2,3,4]:
import numpy as nprng = np.random.default_rng(seed=42)rng.permutation(5)
array([4, 2, 3, 1, 0])
Note that this is equivalent to rng.permutation(np.arange(5)).
Passing in an array
To randomly shuffle an array of numbers:
rng = np.random.default_rng(seed=42)rng.permutation([5,2,6,1])
array([1, 6, 2, 5])
Note that when shuffling one-dimensional arrays, the behaviour is exactly the same as permuted(~).
Setting axis
Consider the following two-dimensional array:
Shuffling the rows
By default, axis=0, which means that the rows are randomly shuffled (in a two-dimensional array):
rng = np.random.default_rng(seed=42)rng.permutation(x) # axis=0
array([[ 8, 9, 10, 11], [ 4, 5, 6, 7], [ 0, 1, 2, 3]])
Shuffling the columns
To randomly shuffle the columns of a two-dimensional array:
rng = np.random.default_rng(seed=42)rng.permutation(x, axis=1)
array([[ 3, 2, 1, 0], [ 7, 6, 5, 4], [11, 10, 9, 8]])
