python最大的k个数_Python笔试题目：求最大的K个数子，解法二，适合小型数据集的情况...

题目：

Givena array of 10,000 random intergers, select the biggest 100 numbers.

1)The order of the result numbers does not matter;

2)Take care about the algorithm performance and big O complexity.

我的解答：

#coding=utf-8

## generate random numbers

from random import randint

# low and high limit of the numbers of the random number

low = -10000000

high = 10000000

# total_number of the numbers

total_number = 10000

# the number of beggest number we need

max_number = 100

# use () for [] will be more efficient ?

numbers = [randint(low,high) for elem in xrange(total_number)]

#print numbers

"""

when the dataset is not large, we still consider the method of "sort and then select"

take quick sort for example, its average complexity is O(N*logN),

then take the K beggest numbers with complexity O(K). So

time complexity is O(N*logN) + O(K) = O(N*logN) in total.

If we only find the biggest K numbers and let the N-K number alone,

the complexity is O(N*K) using part_sorting.

comparing O(N*logN) and O(N*k), we can find that quick sort algorithm is more efficient that the latter one

when k > logN and vice versa.

"""

from math import log

def quick_sort(numbers):

return [] if numbers == [] else quick_sort([y for y in numbers[1:] if y < numbers[0]]) + \

[numbers[0]] + quick_sort([y for y in numbers[1:] if y >= numbers[0]])

def selection_sort_part(numbers):

size = len(numbers)

# the range is [0...max_number-1] ,rather than [0,len(numbers)]. so the selection performs only K times,

for i in range(max_number):

k = i

for j in range(i + 1, size):

#

if numbers[j] > numbers[k]:

k = j

if k is not i:

numbers[i], numbers[k] = numbers[k], numbers[i]

return numbers

def main():

if max_number >= log(total_number,2):

print 'result from quick_sort algorithm:',quick_sort(numbers)[-max_number:]

else:

print 'result from selection_sort_part algorithm:',selection_sort_part(numbers)[:max_number]

if __name__ == '__main__':

main()