Python，Pandas，倒转扩展窗口(Python, Pandas, inverted expanding window)

 让我们考虑以下系列： 
 
s=pd.Series(data=[100, 110, 120, 150, 110, 90, 80, 70, 90, 95])
 
 我遍历每个项目并执行以下操作： 
 
for i in range(len(s)):
    m=max(s[i]-s[i:])
 
 基本上比较每个项目与后面的所有项目，并检索最大值。 
 
 我可以这样做： 
 
r=[]
for i in range(len(s)):
    m=max(s[i]-s[i:])
    r.append(m)

result=pd.Series(r)
 
 但我有这样的感觉，这将是一个滚动窗口可以做的事情？ 我试图让我的代码更清洁一点，并开始阅读和探索滚动功能，但我还没办法，只能用手指。 这个想法看起来像是一个倒转的扩展函数（即它随着数据的传递需要越来越少的数据）？ 
 
 如果有人有关于如何使用滚动/展开窗口重新设置代码格式的想法，请告诉我。 

Lets consider the following Series:
 
s=pd.Series(data=[100, 110, 120, 150, 110, 90, 80, 70, 90, 95])
 
I iterate over each item and perform the following operation:
 
for i in range(len(s)):
    m=max(s[i]-s[i:])
 
Basically comparing each item with all items that follow, and retrieving the max.
 
I can do that with:
 
r=[]
for i in range(len(s)):
    m=max(s[i]-s[i:])
    r.append(m)

result=pd.Series(r)
 
But I had the feeling that this would be something that a rolling window could do? I am trying to make my code a bit cleaner and was just starting reading and exploring about the rolling function, but i cannot yet but my finger on it. This idea seems like a inverted expanding function (i.e. it takes less and less data as it goes down the data)?
 
If someone has ideas about how to reformat the code using rolling / expanding windows, please let me know.

原文：https://stackoverflow.com/questions/47643594