You can use the same indexing approach to swap columns. In this particular case you could avoid the copy by using slice indexing, which returns a view rather than a copy: b = b # invert the row order Note that array indexing always returns a copy rather than a view - there's no way to swap arbitrary rows/columns of an array without generating a copy. Python indexing starts at 0 rather than 1) method (x, axisNone, outNone) Randomly permute x along axis axis. You can perform the swap in a one-liner using integer array indexing: a = np.array(, Universal functions: what, why, and what to do if you want a new one. Furthermore, I need to do an arbitrary number of permutations (more than one). Anatomy of NumPy arrays, and its consequences. axisint, optional Slices of x in this axis are shuffled. np import matplotlib.pyplot as plt import torch import tensorflow as tf print(NumPy:,np. Then, I used reduce () to apply that to m copies of the same array: values combs (np.arange (0, 1, 0.1), 6) for val in. Parameters: xarraylike, at least one-dimensional Array to be shuffled. Transposing and permuting tensors are a common thing to do. First, I created a function that takes two arrays and generate an array with all combinations of values from the two arrays: from numpy import def comb (a, b): c for i in a: for j in b: c.append (r i,j) return c. Unlike shuffle, each slice along the given axis is shuffled independently of the others. That doesn't work for me because the matrices are adjacency matrices (representing graphs), and I need to do the permutations which will give me a graph which is isomorphic with the original graph. method (x, axisNone, outNone) Randomly permute x along axis axis. import numpy as np myarray np.array ( 1, 3, 5, 7, 9) for i in range (5): permutedarray np.random. numpy.shuffle and numpy.permutation seem to permute only the rows of the matrix (not the columns at the same time). Put your array as an argument of random permutation function. Now, an incredibly naive (and memory costly) way of doing so might be: a2 = deepcopy(a1)īut, I would like to know if there is something more efficient that does this. Assuming that I have the following matrix/array: array(,Īnd I want to apply the following permutation: 1 -> 5

Numpy permute