为什么在调用完全执行此操作的函数时必须重做 try-except 语句？

Question

I have the following problem: I have specified a method for a given class which handles two different Errors with the same user-defined exception (see the method below).

def get_route(self, source, target):

    try:
        n = nx.shortest_path(self.graph, source, target)
        return n

    except (nx.NetworkXNoPath, nx.NodeNotFound):
        raise NoSuchFlight

So now, I want to further define a function which calls the above method. I have noticed that the code works when I redo a try-except statement I do not really understand (see below).

def get_airlines_per_route(airlines_dict, source, target, max_intermediate_stops):
    lst = []

    for airline in airlines_dict.values():

        try:
            path = airline.get_route(source, target)
            if (len(path) - 2) <= max_intermediate_stops:
                lst.append(airline)

        except:
            NoSuchFlight

    return lst

However, I have no idea why this works. Anyway is my code correct? And moreover, how would you guys have tackled this problem?

Thanks for any advice.

Update:

The user-defined exception:

class NoSuchFlight(Exception):
    pass

Intended way:

def get_airlines_per_route(airlines_dict, source, target, max_intermediate_stops):
    lst = []

    for airline in airlines_dict.values():

        path = airline.get_route(source, target)
        if (len(path) - 2) <= max_intermediate_stops:
            lst.append(airline)

    return lst

Answer 1

the second try except will catch all the Exceptions this why your code is working, I suggest to use:

try:
    path = airline.get_route(source, target)
    if (len(path) - 2) <= max_intermediate_stops:
        lst.append(airline)

except NoSuchFlight:
    # log or something else

Answer 2

It works but you should not have code like that.

In the inner function you raise a custom exception in abnormal conditions. This part is correct, even if it is more common to encapsulate the original error for eventual further analysis. So I would probably use:

except (nx.NetworkXNoPath, nx.NodeNotFound) as e:
    raise NoSuchFlight(e)

The outer code is really weird: except: will catch any exception. This is bad practice because you should not catch exception that you do not intend to process. For example, an IO error because of a physical disk problem should probably not be ignored.

And the following line, NoSuchFlight is close to a no-op: it just names an exception. You can try to type Exception in a python interpretor and it will just display the class name:

>>> Exception
<class 'Exception'>

without raising anything...

If you just want to ignore erroneous routes, you should use:

def get_airlines_per_route(airlines_dict, source, target, max_intermediate_stops):
    lst = []

    for airline in airlines_dict.values():
       try:
           path = airline.get_route(source, target)
           if (len(path) - 2) <= max_intermediate_stops:
               lst.append(airline)
       catch  NoSuchFlight:          # just skip in case of a NoSuchFlight
           pass

    return lst