Question

此实现使用链接列表而不是使用内置列表。有没有人使用哪个版本的性能更好？

class Stack:
    top = ''
    def __init__(self,data=None,next=None):
        self.data = data
        self.next = next
    def pop(self):
        if self.top != None:
            item = self.top.getvalue()
            self.top = self.top.next
            return item
        else:
            return 
    def push(self,data):
        t = Stack(data)
        t.next = self.top
        self.top = t
    def peek(self):
        return self.top.getvalue()
    def getvalue(self):
        return self.data

s = Stack()
s.push('bottom')
s.push('middle')
s.push('top')
popped = s.pop()
print(popped)
top = s.peek()
print(top)

输出：
顶部
中间

Answer 1

基于内置列表的堆栈类将具有更多更好的性能，例如

class StackEmptyError(Exception):
    pass

class ListBasedStack(list):
    push = list.append

    def pop(self):
        try:
            return super(ListBasedStack, self).pop()
        except IndexError:
            raise StackEmptyError

    def peek(self):
        try:
            return self[-1]
        except IndexError:
            raise StackEmptyError

使用timeit模块测量执行时间：

>>> import timeit
>>> stmt = '''
s = Stack()
s.push('bottom')
s.push('middle')
s.push('top')
s.peek()
s.pop()'''

# test built-in list stack class
>>> timeit.repeat(stmt, 'from __main__ import ListBasedStack as Stack')
[1.4590871334075928, 1.4085769653320312, 1.3971672058105469]

# test OP stack class
>>> timeit.repeat(stmt, 'from __main__ import Stack')
[5.018421173095703, 4.971158981323242, 4.990453004837036]

基于内置列表的类比链接列表快约3.5倍。

而且，如果您对于为IndexError操作提出Stack例外情况并不挑剔，那么它会更快（大约快5倍）：

class MinimalistStack(list):
    push = list.append
    def peek(self):
        return self[-1]

>>> timeit.repeat(stmt, 'from __main__ import MinimalistStack as Stack')
[0.9379069805145264, 0.9144589900970459, 0.9160430431365967]

在Python中创建堆栈类

1 个答案: