Creating a list of strings defined by similar items in a `numpy.array`

Question

Given:

list of strings:

colors = ['red', 'blue', 'green', 'grey', 'black', 'yellow', 'brown', 'orange', 'magenta', 'cyan', 'teal', 'turquise']

a, a numpy.array with a.shape[0] <= len(colors) that contains some numeric (repeating) values.

Requirement:

Producing a list l, of str objects that will match the elements in a. I.e. for each element in a, there will be the same str object in l, as in shown in Example 1.

Example 1:

Given:

import numpy as np


colors = ['red', 'blue', 'green', 'grey', 'black', 'yellow', 'brown', 'orange', 'magenta', 'cyan', 'teal', 'turquise']

a = np.array([0, 1, 5, 7, 1, 5, 7, 7, 0, 0])

the desired output should be:

Out: ['red', 'blue', 'green', 'grey', 'blue', 'green', 'grey', 'grey', 'red', 'red']

My solution:

I can think of two ways to do it as shown in Code 1 and Code 2.

Code 1:

import numpy as np

colors = ['red', 'blue', 'green', 'grey', 'black', 'yellow', 'brown', 'orange', 'magenta', 'cyan', 'teal', 'turquise']
a = np.array([0, 1, 5, 7, 1, 5, 7, 7, 0, 0])

num_2_color_dict = {}
for num in np.unique(a):
    num_2_color_dict[num] = colors[np.min(np.where(a==num))]
    
color_by_num = []
for num in a:
    color_by_num.append(num_2_color_dict[num])
color_by_num

Out: ['red', 'blue', 'green', 'grey', 'blue', 'green', 'grey', 'grey', 'red', 'red']

Code 2:

import numpy as np

colors = ['red', 'blue', 'green', 'grey', 'black', 'yellow', 'brown', 'orange', 'magenta', 'cyan', 'teal', 'turquise']
a = np.array([0, 1, 5, 7, 1, 5, 7, 7, 0, 0])

colors_by_num = [colors[np.min(np.where(np.unique(a)==num))] for num in a]
colors_by_num

Question:

I'd like to be able to achieve the same result without using loops, and preferably with some numpy array manipulation technique, that will shorten the code.

Thanks in advance.

MichaelJanz · Accepted Answer · 2020-08-10 11:38:09Z

4

As Josef states in his answer, numpy.unique has the option to return integer encodings, which is what you want. So the following piece of code should help you out:

np.take(colors, np.unique(a,return_inverse=True)[1])

With setting return_inverse=True, you get the integer encodings, as second return parameter. Numpy.take() gives you the items for the indices.

answered Aug 10, 2020 at 11:38

MichaelJanz

1,8252 gold badges10 silver badges24 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Hammad Ahmed · Accepted Answer · 2020-08-10 11:42:04Z

1

colors = np.array(['red', 'blue', 'green', 'grey', 'black', 'yellow', 'brown', 'orange', 'magenta', 'cyan', 'teal', 'turquise'])
a      = np.array([0, 1, 5, 7, 1, 5, 7, 7, 0, 0])

_, i = np.where(a[:, None] == np.unique(a))

colors[i]


>>> array(['red', 'blue', 'green', 'grey', 'blue', 'green', 'grey', 'grey', 'red', 'red'])

answered Aug 10, 2020 at 11:42

Hammad Ahmed

8858 silver badges17 bronze badges

Collectives™ on Stack Overflow

Creating a list of strings defined by similar items in a `numpy.array`

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related