I have a 2 numpy array something like this
a = [array([ 0.1, 0.1, 0.1]), array([ 0.2, 0.2, 0.2])]
b = [0 0 0 1]
What I want is something like this --
c = [[0.1, 0.1, 0.1],[0.1, 0.1, 0.1],[0.1, 0.1, 0.1],[0.2, 0.2, 0.2]]
i.e. elements of a based on index of b.
Is there a way I can achieve this using numpy and vectorization i.e. without looping over the values?
If you store a as a two-dimensional numpy array:
>>> a = np.array([[0.1, 0.1, 0.1], [0.2, 0.2, 0.2]])
# result: array([[ 0.1, 0.1, 0.1],
# [ 0.2, 0.2, 0.2]])
or even convert a to a numpy array via a = np.array(a),
then you can use the list b to access the elements as desired:
>>> b = [0,0,0,1]
>>> print(a[b])
array([[ 0.1, 0.1, 0.1],
[ 0.1, 0.1, 0.1],
[ 0.1, 0.1, 0.1],
[ 0.2, 0.2, 0.2]])
and if you need a list as output then use tolist() method of the numpy arrays:
>>> (np.asarray(a)[b]).tolist()
[[0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.2, 0.2, 0.2]]
list comprehension
[a[x].tolist() for x in b]
import numpy
a = [numpy.array([ 0.1, 0.1, 0.1]), numpy.array([ 0.2, 0.2, 0.2])]
b = [0, 0, 0, 1]
Alternative 1:
print([a[x].tolist() for x in b])
Output:
[[0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.2, 0.2, 0.2]]
Alternative 2:
print(numpy.array(a)[b])
Output:
[[ 0.1 0.1 0.1]
[ 0.1 0.1 0.1]
[ 0.1 0.1 0.1]
[ 0.2 0.2 0.2]]
Alternative 3:
print(list(map(lambda i: a[i], b)))
Output:
[array([ 0.1, 0.1, 0.1]), array([ 0.1, 0.1, 0.1]), array([ 0.1, 0.1, 0.1]), array([ 0.2, 0.2, 0.2])]
Alternative 4:
from operator import itemgetter
print(list(itemgetter(*b)(a)))
Output:
[array([ 0.1, 0.1, 0.1]), array([ 0.1, 0.1, 0.1]), array([ 0.1, 0.1, 0.1]), array([ 0.2, 0.2, 0.2])]
Using numpy
If you want using numpy then:
print([a[i].tolist() for i in b])
Without using numpy :
import numpy as np
a = np.array([[0.1, 0.1, 0.1], [0.2, 0.2, 0.2]])
b = [0,0,0,1]
print([value_1.tolist() for value in b for index,value_1 in enumerate(a) if index==value])
above list comprehension is same as :
final=[]
for value in b:
for index,value_1 in enumerate(a):
if index==value:
final.append(value_1.tolist())
print(final)
output:
[[0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.1, 0.1, 0.1], [0.2, 0.2, 0.2]]
