美文网首页
Python 八大排序算法速度比较

Python 八大排序算法速度比较

作者: 依天立业 | 来源:发表于2018-09-18 17:22 被阅读0次

    Python 八大排序算法速度比较

    这篇文章并不是介绍排序算法原理的,纯粹是想比较一下各种排序算法在真实场景下的运行速度。

    算法由 Python 实现,用到了一些语法糖,可能会和其他语言有些区别,仅当参考就好。

    测试的数据是自动生成的,以数组形式保存到文件中,保证数据源的一致性。

    排序算法

    image

    直接插入排序

    • 时间复杂度:O(n²)
    • 空间复杂度:O(1)
    • 稳定性:稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def insert_sort(array): for i in range(len(array)): for j in range(i): if array[i] < array[j]:
    array.insert(j, array.pop(i)) break
    return array</pre>

    希尔排序

    • 时间复杂度:O(n)
    • 空间复杂度:O(n√n)
    • 稳定性:不稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def shell_sort(array):
    gap = len(array) while gap > 1:
    gap = gap // 2
    for i in range(gap, len(array)): for j in range(i % gap, i, gap): if array[i] < array[j]:
    array[i], array[j] = array[j], array[i] return array</pre>

    简单选择排序

    • 时间复杂度:O(n²)
    • 空间复杂度:O(1)
    • 稳定性:不稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def select_sort(array): for i in range(len(array)):
    x = i # min index
    for j in range(i, len(array)): if array[j] < array[x]:
    x = j
    array[i], array[x] = array[x], array[i] return array</pre>

    堆排序

    • 时间复杂度:O(nlog₂n)
    • 空间复杂度:O(1)
    • 稳定性:不稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def heap_sort(array): def heap_adjust(parent):
    child = 2 * parent + 1 # left child
    while child < len(heap): if child + 1 < len(heap): if heap[child + 1] > heap[child]:
    child += 1 # right child
    if heap[parent] >= heap[child]: break heap[parent], heap[child] =
    heap[child], heap[parent]
    parent, child = child, 2 * child + 1 heap, array = array.copy(), [] for i in range(len(heap) // 2, -1, -1):
    heap_adjust(i) while len(heap) != 0:
    heap[0], heap[-1] = heap[-1], heap[0]
    array.insert(0, heap.pop())
    heap_adjust(0) return array</pre>

    冒泡排序

    • 时间复杂度:O(n²)
    • 空间复杂度:O(1)
    • 稳定性:稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def bubble_sort(array): for i in range(len(array)): for j in range(i, len(array)): if array[i] > array[j]:
    array[i], array[j] = array[j], array[i] return array</pre>

    快速排序

    • 时间复杂度:O(nlog₂n)
    • 空间复杂度:O(nlog₂n)
    • 稳定性:不稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def quick_sort(array): def recursive(begin, end): if begin > end: return l, r = begin, end
    pivot = array[l] while l < r: while l < r and array[r] > pivot:
    r -= 1
    while l < r and array[l] <= pivot:
    l += 1 array[l], array[r] = array[r], array[l]
    array[l], array[begin] = pivot, array[l]
    recursive(begin, l - 1)
    recursive(r + 1, end)

    recursive(0, len(array) - 1) return array</pre>
    

    归并排序

    • 时间复杂度:O(nlog₂n)
    • 空间复杂度:O(1)
    • 稳定性:稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def merge_sort(array): def merge_arr(arr_l, arr_r):
    array = [] while len(arr_l) and len(arr_r): if arr_l[0] <= arr_r[0]:
    array.append(arr_l.pop(0)) elif arr_l[0] > arr_r[0]:
    array.append(arr_r.pop(0)) if len(arr_l) != 0:
    array += arr_l elif len(arr_r) != 0:
    array += arr_r return array def recursive(array): if len(array) == 1: return array
    mid = len(array) // 2 arr_l = recursive(array[:mid])
    arr_r = recursive(array[mid:]) return merge_arr(arr_l, arr_r) return recursive(array)</pre>

    基数排序

    • 时间复杂度:O(d(r+n))
    • 空间复杂度:O(rd+n)
    • 稳定性:稳定

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">def radix_sort(array):
    bucket, digit = [[]], 0 while len(bucket[0]) != len(array):
    bucket = [[], [], [], [], [], [], [], [], [], []] for i in range(len(array)):
    num = (array[i] // 10 ** digit) % 10 bucket[num].append(array[i])
    array.clear() for i in range(len(bucket)):
    array += bucket[i]
    digit += 1
    return array</pre>

    速度比较

    image

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">from random import random from json import dumps, loads # 生成随机数文件
    def dump_random_array(file='numbers.json', size=10 ** 4):
    fo = open(file, 'w', 1024)
    numlst = list() for i in range(size):
    numlst.append(int(random() * 10 ** 10))
    fo.write(dumps(numlst))
    fo.close() # 加载随机数列表
    def load_random_array(file='numbers.json'):
    fo = open(file, 'r', 1024) try:
    numlst = fo.read() finally:
    fo.close() return loads(numlst)</pre>

    image

    显示执行时间

    如果数据量特别大,采用分治算法的快速排序和归并排序,可能会出现递归层次超出限制的错误。

    解决办法:导入 sys 模块(import sys),设置最大递归次数(sys.setrecursionlimit(10 ** 8))。

    <pre style="margin: 0px; padding: 0px; white-space: pre-wrap; word-wrap: break-word; font-family: "Source Code Pro", Consolas, Menlo, Monaco, "Courier New", monospace !important; font-size: 0.8rem !important; background: none !important; line-height: 1.2rem !important;">@exectime def bubble_sort(array): for i in range(len(array)): for j in range(i, len(array)): if array[i] > array[j]:
    array[i], array[j] = array[j], array[i] return array

    array = load_random_array() print(bubble_sort(array) == sorted(array))</pre>

    ↑ 示例:测试直接插入排序算法的运行时间,@exectime 为执行时间装饰器。

    算法执行时间

    image

    算法速度比较

    image image

    相关文章

      网友评论

          本文标题:Python 八大排序算法速度比较

          本文链接:https://www.haomeiwen.com/subject/wazpnftx.html