CODEWITHSUNDEEP
pythonnumpy
piRSquared
piRSquared

Reputation: 294278

can we deduce an argsort array from numpy's unique function?

From the docs

Returns the sorted unique elements of an array. There are three optional outputs in addition to the unique elements: the indices of the input array that give the unique values, the indices of the unique array that reconstruct the input array, and the number of times each unique value comes up in the input array.

Therefore, numpy.unique must be performing a sorting algorithm. Is there a way to deduce the sorted-ness from the outputs of numpy.unique. Asked another way, can we derive the same results of an argsort purely from the output of numpy.unique so I don't have to sort twice.

Consider arrays a and b as sample data.

import numpy as np
from string import ascii_letters

np.random.seed([3,1415])
n = 10000
a = np.random.randint(100, size=n)
b = np.random.choice(list(ascii_letters), n)

Upvotes: 1

Views: 460

Answers (1)

piRSquared
piRSquared

Reputation: 294278

I just wrote my own unique function

def unique(a):
    s = a.argsort()
    a_ = a[s]
    return a_[np.append(True, (a_[1:] != a_[:-1]))], s

timing
@hpaulj is correct. Notice that my custom unique is just as fast as np.unique when np.unique is asked to return an inverse array. Otherwise, when it just returns unique values, it is a bit faster.

enter image description here

Upvotes: 1

Related Questions