即使最后一个命令（但不是第一个命令）引发异常，try块也会给出输出

Question

我使用try / except来逐行读取文件时捕获问题。 try块包含一系列操作，最后一个操作通常是异常的原因。令人惊讶的是，我注意到即使引发异常，所有先前的操作都在try块中执行。尝试将我创建的字典转换为数据框时，这是一个问题，因为列表的长度不相等。

此代码会产生问题：

d = {'dates':[],'states':[], 'longitude':[], 'latitude':[], 'tweet_ids':[], 'user_ids':[], 'source':[]}
for file in f:
    print("Processing file "+file)
    t1 = file.split('/')[-1].split("_")
    date = t1[0]
    state_code = t1[1]
    state = list(states_ref.loc[states_ref.code==state_code]['abbr'])[0]

    collection = JsonCollection(file)
    counter = 0
    for tweet in collection.get_iterator():
        counter += 1
        try:

            d['dates'].append(date)
            d['states'].append(state)
            t2 = tweet_parser.get_entity_field('geo', tweet)
            if t2 == None:
                d['longitude'].append(t2)
                d['latitude'].append(t2)
            else:
                d['longitude'].append(t2['coordinates'][1])
                d['latitude'].append(t2['coordinates'][0])


#note: the 3 lines bellow are the ones that can raise an exception 
            temp = tweet_parser.get_entity_field('source', tweet)
            t5 =  re.findall(r'>(.*?)<', temp)[0]
            d['source'].append(t5)

        except:
            c += 1
            print("Tweet {} in file {} had a problem and got skipped".format(counter, file))
            print("This is a total  of {} tweets I am missing from the {} archive I process.".format(c, sys.argv[1]))
            next

tab = pd.DataFrame.from_dict(d)

我通过移动易于在顶部给出错误的操作来解决问题，但我想更好地理解为什么try / except表现得像这样。有任何想法吗？

此代码有效：

d = {'dates':[],'states':[], 'longitude':[], 'latitude':[], 'tweet_ids':[], 'user_ids':[], 'source':[]}
for file in f:
    print("Processing file "+file)
    t1 = file.split('/')[-1].split("_")
    date = t1[0]
    state_code = t1[1]
    state = list(states_ref.loc[states_ref.code==state_code]['abbr'])[0]

    collection = JsonCollection(file)
    counter = 0
    for tweet in collection.get_iterator():
        counter += 1
        try:
            #note: the 3 lines bellow are the ones that can raise an exception 
temp = tweet_parser.get_entity_field('source', tweet)
            t5 =  re.findall(r'>(.*?)<', temp)[0]
            d['source'].append(t5)

            d['dates'].append(date)
            d['states'].append(state)
            t2 = tweet_parser.get_entity_field('geo', tweet)
            if t2 == None:
                d['longitude'].append(t2)
                d['latitude'].append(t2)
            else:
                d['longitude'].append(t2['coordinates'][1])
                d['latitude'].append(t2['coordinates'][0])
        except:
            c += 1
            print("Tweet {} in file {} had a problem and got skipped".format(counter, file))
            print("This is a total  of {} tweets I am missing from the {} archive I process.".format(c, sys.argv[1]))
            next

tab = pd.DataFrame.from_dict(d)

Answer 1

在附加到目标对象之前，您始终可以使用时态对象来保存函数的输出。这样，如果某些内容失败，它会在将数据放入目标对象之前引发异常。

try:
    #Put all data into a temporal Dictionary
    #Can raise an exception here
    temp = tweet_parser.get_entity_field('source', tweet)
    t2 = tweet_parser.get_entity_field('geo', tweet)
    tempDictionary = {
        "source"    : re.findall(r'>(.*?)<', temp)[0],
        "latitude"  : None if (t2 is None) else t2['coordinates'][1],
        "longitude" : None if (t2 is None) else t2['coordinates'][0]
    }
    #Append data from temporal Dictionary
    d['source'].append(tempDictionary['source'])
    d['latitude'].append(tempDictionary['latitude'])
    d['longitude'].append(tempDictionary['longitude'])
    d['dates'].append(date)
    d['states'].append(state)
except:
    c += 1
    print("Tweet {} in file {} had a problem and got skipped".format(counter, file))
    print("This is a total  of {} tweets I am missing from the {} archive I process.".format(c, sys.argv[1]))