如何从计算中排除列表的最大数量

Question

I have a list of lists with floats and want to get the mean of all the items of a list, but always leave the biggest number out. The one float that was not included should then be subtracted from the mean since this would be the biggest discrepancy. Then I want to create a list to put the discrepancy in.

original_list = [[0.5, 1.4, 2.1, 5.2], [2.3, 3.2, 5.3, 2.1], [1.3, 3.1, 2.1, 5.3]]
result_list = [[-3.87], [-2.77], [-3.13]] 

I don't know how to exclude the biggest number from calculations because it's always in another position.

Answer 1

We can loop over the list, sort each sublist and then perform the calculation, appending the result to a result list.

import numpy as np
original_list = [[0.5, 1.4, 2.1, 5.2], [2.3, 3.2, 5.3, 2.1], [1.3, 3.1, 2.1, 5.3]]

rv = []

for x in original_list:
    l = sorted(x)
    rv.append([np.mean(l[:-1])-l[-1]])

Output:

>>> rv
[[-3.866666666666667], [-2.766666666666666], [-3.1333333333333333]]

To round the values use rv.append([round(np.mean(l[:-1])-l[-1], 2)]) :

>>> rv
[[-3.87], [-2.77], [-3.13]]

Answer 2

Pure python solutions

original_list = [[0.5, 1.4, 2.1, 5.2], [2.3, 3.2, 5.3, 2.1], [1.3, 3.1, 2.1, 5.3]]

def mean_no_max(l):
    s, m = 0, float('-inf')
    for i in l:
        s += i
        m = m if m > i else i
    return [(s - m) / (len(l) - 1) - m]

print(list(map(mean_no_max, original_list)))  # -> [[-3.86], [-2.76], [-3.13]]

---

Fun approach with single iteration over l using max() and sum()

def mean_no_max(l):
    s = [0]
    def update_sum(x):
        s[0] += x
        return x
    max_value = max(l, key=update_sum)
    return (s[0] - max_value) / (len(l) - 1) - max_value

---

def mean_no_max(l):
    class Adder:
        def __init__(self, num=0, max_value=float('-inf')):
            self.num = num
            self.max_value = max_value
        def __add__(self, other):
            self.num += other
            if other > self.max_value:
                self.max_value = other
            return self
    adder = sum(l, Adder())
    return (adder.num - adder.max_value) / (len(l) - 1) - adder.max_value

Answer 3

as comprehension

from numpy import mean

original_list = [[0.5, 1.4, 2.1, 5.2], [2.3, 3.2, 5.3, 2.1], [1.3, 3.1, 2.1, 5.3]]

result_list = [[round(mean([v for v in list_ if v !=max(list_)])-max(list_), 2)] for list_ in original_list]
print(result_list)

Answer 4

One line:

import numpy as np
res = [[np.mean(sorted(i)[:-1])-sorted(i)[-1]] for i in original_list]
print(res)

Answer 5

you can get first a list with all the max values from your nested lists and then filter the nested lists by max values

max_values = [max(l) for l in original_list]
with_no_max = [[e for e in l if e < m] for l, m in zip(original_list, max_values)]

print(max_values)
print(with_no_max)

output:

[5.2, 5.3, 5.3]
[[0.5, 1.4, 2.1], [2.3, 3.2, 2.1], [1.3, 3.1, 2.1]]