如何实现一个函数来撤消修改列表的最后一个输入？

Question

So I recently made a program that adds, removes (either just plain removes or by a user input [ie if an item is less than a user input then it gets deleted and the such]) and modifies packages (dictionaries containing beginning and end dates, destination and price) in a list

When I implemented the remove/modify functions, I also made sure to append the original item to a new list so that when I had to call the undo function it'd just append the original item to the original list again, however, the function just deletes the item that was in the list after the remove/modify calls and just appends the item that got removed/modified

What I tried to do (my train of thought) was that I wanted to create a new list called "lst" where I stock the items that were about to be removed, and I would essentially use it in the undo function (basically the function that you call undo list). In the undo function I enumerate thru l and lst, and I tried to append the item from the lst list. Now I see how I messed up given that the index will probably not be the same and I caught an IndexError so I resorted to popping the previous item

the code: https://pastebin.com/taZAjhQf the function that I was talking about:

def undo(l,lst):
    for el,pachet in enumerate(l):
        l.append(lst[el])
        l.pop()

if you could at least give me a hint or tell me if I'm on the right track...

Answer 1

My suggestion is to maintain a log of the operations that took place (similar to what accountants do with their bookkeeping). That way you're storing the operations that took place rather than the actual state. You can construct the final state by following your log and applying the operations that took place in order. For undoing, you can pop off the last operation and reconstruct the state (I'll leave it to you to think of how you might optimise that)

This technique is quite popular these days, especially in systems that make use of Kafka. For more ideas into this technique, you can check out this free ebook from Jay Kreps and there's also a nice article on it.