I am trying to find the N biggest values from a list, and then print out their positions in the list.
If I would only focus on the max. value, it would look like this:
>>>>fr = [8,4,1,1,12]
>>>>print fr.index(max(fr))
4
However, my aim is to get an output like: 4,0,1 if it were for n=3. The 0 here shows the position of the second biggest value. REMEMBER, I am not interested in the value, but in their position!
Use heapq.nlargest with key = fr.__getitem__:
>>> import heapq
>>> fr = [8,4,1,1,12]
>>> heapq.nlargest(3, xrange(len(fr)), key=fr.__getitem__)
[4, 0, 1]
If you want the values itself, then:
>>> heapq.nlargest(3, fr)
[12, 8, 4]
xrange(len(fr)) based on the key i.e the value at those indices in fr, and return the 3 largest value. It is not exactly sorting but using heapq algorithm to get the largest items.Another way is:
[fr.index(x) for x in sorted(fr, reverse=True)[:3]]
When we compare speed of both of them...
import heapq
fr = [8, 4, 1, 1, 12]
def method_one():
for i in xrange(10000):
res = [fr.index(x) for x in sorted(fr, reverse=True)[:3]]
def method_two():
for i in xrange(10000):
heapq.nlargest(3, xrange(len(fr)), key=fr.__getitem__)
if __name__ == '__main__':
import timeit
print timeit.repeat(stmt='method_one()',
setup='from __main__ import method_one',
number=100)
print timeit.repeat(stmt='method_two()',
setup='from __main__ import method_two',
number=100)
we get:
[1.1253619194030762, 1.1268768310546875, 1.128382921218872]
[2.5129621028900146, 2.529547929763794, 2.492828130722046]
[fr.index(x) for x in sorted(fr, reverse=True)][:3], whereas this will get the first 3 elements.fr = [8,8,1,2,12] and n = 4, then result=[4, 0, 0, 3] which may not be what the OP asked for. The heap solution gives the correct [4,0,1,3] indices thoughThis simplest way is to just do this
>>> fr = [8,4,1,1,12]
>>> n = 3
>>> result = [fr.index(i) for i in sorted(fr, reverse=True)][:n]
>>> print(result)
[4, 0, 1]
No libraries and dependancies.
it's always easy to convert the list to numpy array and deal with it.
import numpy as np
arr = np.array([10, 5, 8, 15, 15, 12])
sorted_indices = np.argsort(arr)[::-1]
largest_indices = sorted_indices[:3]
print(largest_indices)
Output: [4 3 5]
select = 4
list_val = [5.0, 0, 0, 0, 0, 5.0]
result = [list_val.index(i) for i in sorted(list_val, reverse=True)][:select]
print(result)
#expected: [0, 5, 1, 2]
#output: [0, 0, 1, 1]
This method won't always work.
0in input?0is the position of the2ndlargest number in the list.