Treverse list of tuples to compare and report min, max-CodePudding

My previous question was not understood, so I rephrase and post this one. I have a list of tuple for (class, n_class_examples) like this:

my_list = (0, 126), (1, 192), (2, 330), (3, 952) ]

So I am interested in generating a function, that takes in such a list, compare each tuple against all others, and in each case reports which class has smaller number of samples (min_class), and which has the larger number of samples (max_class).

def get_min_max_class(current_list):
    for tn, tn 1: # tn -> 1-tuple, tn 1 any other tuple not tn
        if tn[1] < tn 1[1]
            smaller_class = tn[0]
            larger_class = tn 1[0]
        smaller_class = tn 1[0]
        larger_class = tn[0]
    return # smaller, larger of the 2 compared in each case

So that:

get_min_max_class(my_list)
# would perform the comparison like so:
(0, 126) v (1, 192) -> min_class = 0, max_class = 1 # in this case
(0, 126) v (2, 330) -> min_class = 0, max_class = 2 # and in this case
(0, 126) v (3, 952) -> min_class = 0, max_class = 3 # and here ..
(1, 192) v (2, 330) -> min_class = 1, max_class = 2 # ...
(1, 192) v (3, 952) -> min_class = 1, max_class = 3
(2, 330) v (3, 952) -> min_class = 2, max_class = 3

Forgive my definition of function, but I want the function to iteratively compare those items, each time, report which is larger and which is smaller.

CodePudding user response：

Iterate over the list of pairs generated by itertools.combintions, the process each pair individually using min and max.

from itertools import combinations
from operator import itemgetter

first = itemgetter(0)
second = itemgetter(1)

def get_min_max_class(current_list):
    for pair in combinations(current_list, 2):
        p0, p1 = pair
        min_class = first(min(pair, key=second))
        max_class = first(max(pair, key=second))
        print(f'{p0} v {p1} -> min_class = {min_class}, max_class = {max_class}')

get_min_max_class(my_list)

If you want to return a list of results, rather than simply printing a report, you'll have to define what exactly you want to return.

CodePudding user response：

Python's sorted(), min(), and max() functions take as second argument a 'key' that let's you specify how to calculate the sorting for different objects, using lambda functions. In this case, you want to sort the tuples based on the value of the second element and return the corresponding first value.

So, if I wanted the 'max' in your case I would do:

max_class = max(my_list, key=lambda x: x[1])[-1]

The lambda expression, if it's new to you, is saying "sort the item x in the list based on whatever you find at x[1]". Then, take the final element of the sorted list to get the class with the most samples, or whatever it was exactly.

I hope that helps!