python缩短嵌入式try-except块

Question

I've been using try-except inside another try-except , when I try to open a url which may or may not lack an 'http://' header.

But the code looks messy. I'm wondering if python has some nicer ways to deal with such needs. Actually I've read the 'with' keyword.. Which somehow, I feel, will make the code harder to read...

I've got links to open, but the links may have something missing. For example the url is like. either xxx.com or http : / / xxx . com. And I have to test http:// and url, http:// plus www. plus url, url to find the right one.

And if none works, I must record the reason why, and go on with the next link:

    for link in links:
        url = link
        if not url.startswith('http'):
            try:
                url1 = r'http://'+link
                res = requests.get(url)
            except Exception as e:
                try:
                    url2 = r'http://www.'+link
                    res = requests.get(url)
                except Exception as err:
                    self.print_error(*[url1, modify(e.message)])
                    self.print_error(*[url2, modify(err.message)])
                    self.error_log(*[url1, modify(e.message)])
                    self.error_log(*[url2, modify(err.message)])

        else:
            try:
                res = request.get(url)
            except Exception as e:
                self.print_error(*[url, modify(e.message)])
                self.error_log(*[url, modify(e.message)])

Answer 1

First assemble a list of potential URLs in order of priority:

potential_urls = [link, 'http://' + link, 'http://www.' + link]

Also keep a list of errors you've encountered:

errors_encountered = []

Then go through the list, break ing if it works.

res = None
for url in potential_urls:
    try:
        res = requests.get(url)
    except Exception as err:
        errors_encountered.append(err)
    else:
        break

If none of the attempts work, we'll leave res set to None , which we can detect and log as desired:

if res is None:
    for url, err in zip(potential_urls, errors_encountered):
        self.print_error(url, modify(err.message))
    for url, err in zip(potential_urls, errors_encountered):
        self.error_log(url, modify(err.message))

Otherwise, it succeeded, and you can use res as usual.

Answer 2

Here is one other possibility, which simply reduces the nesting by continuing through the loop on success.

for link in links:
    url = link
    if not url.startswith('http'):
        try:
            url1 = r'http://'+link
            res = requests.get(url)
            continue # we succeeded
        except Exception as e:
            pass

        try:
            url2 = r'http://www.'+link
            res = requests.get(url)
        except Exception as err:
            self.print_error(*[url1, modify(e.message)])
            self.print_error(*[url2, modify(err.message)])
            self.error_log(*[url1, modify(e.message)])
            self.error_log(*[url2, modify(err.message)])

Answer 3

Similar to @icktoofay's answer, but working from the perspective that you probably only need to work if the link succeeded or not.

Define a function that does possible attempts and either returns a requests object, or after attempting all possibilities, raises an exception.

def try_get(link):
    for url in (link, 'http://' + link, 'http://www.' + link):
        try:
            return requests.get(url)
        except Exception:
            pass
    raise ValueError('no url was suitable', link)

Then loop over your urls, and either use res or do something suitable with the exception.

for url in list_of_urls:
    try:
        res = try_get(url)
    except ValueError as e:
        print e