Is np.argpartition giving me the wrong results?-CodePudding

Take the following code:

import numpy as np
one_dim = np.array([2, 3, 1, 5, 4])
partitioned = np.argpartition(one_dim, 0)
print(f'Unpartitioned array: {one_dim}')
print(f'Partitioned array index: {partitioned}')
print(f'Partitioned array: {one_dim[partitioned]}')

The following output results:

Unpartitioned array: [2 3 1 5 4]
Partitioned array index: [2 1 0 3 4]
Partitioned array: [1 3 2 5 4]

The output for the partitioned array should be [1 2 3 5 4]. How is three on the left side of two? It seems to me the function is making an error or am I missing something?

CodePudding user response：

The second argument is which index will be in sorted position after partitioning, so it is correct that index 0 of the partition (element value 1) is in sorted position, and everything to the right is greater.