##References Using those references:

• rolling median implementations benchmark
• FAST RUNNING MEDIAN USING AN INDEXABLE SKIPLIST
• Calculate the median of a billion numbers
• Find running median from a stream of integers

##Code A code was written to calculate the Median with the SortedContainers library:

from itertools import islice
from sortedcontainers import SortedList
import random
import time

start_time = time.time()


class Median(object):

    def __init__(self, iterable):
        self._iterable = islice(iterable, None)
        self._sortedlist = SortedList(self._iterable)

    def __iter__(self):
        self_sortedlist = self._sortedlist
        # print(self_sortedlist)
        length = len(self_sortedlist)
        half = length // 2
        if length % 2 == 0:
            yield (self_sortedlist[half] + self_sortedlist[half - 1]) // 2
        elif length % 2 == 1:
            yield self_sortedlist[half]


def main():
    m, n = 1000, 1500000
    data = [random.randrange(m) for i in range(n)]
    # print("Random Data: ", data)
    result = list(Median(data))
    print("Result: ", result)


if __name__ == "__main__":

    main()
    print("--- %s seconds ---" % (time.time() - start_time))

##Explanation

###Random Number Generator The following code generates data within the Range m and quantity n.

m, n = 1000, 15000000
data = [random.randrange(m) for i in range(n)]

###Median

The Median class sorts the list of numbers and if n is odd, returns the middle item with yield self_sortedlist[half]. Or if n is even, returns the mean of two middle itens of the list with yield (self_sortedlist[half] + self_sortedlist[half - 1]) // 2

##Question

How do I improve the code performance? Because for a large list (100 milion), it takes --- 186.7168517112732 seconds--- on my computer.