Python在Try/Except循环中触发KeyError问题排查

问题场景 #

运行以下代码时，当接口响应的JSON不包含code键，系统会抛出key not found 'code'的错误，且部署在Heroku上的Flask应用直接崩溃，没有进入except TypeError代码块。

def create_hec_kw(self, kw):
    print(f'Creating Keyword {kw}')
    data = {'name': kw, 'slug': kw.lower().replace(' ', '-')}
    response = requests.post(self.create_url('tags'), headers=self.get_headers(), json=data)
    # created_kw_id = response.json()['data']['term_id'] if response.json()['code'] else response.json()['id']
    print(f'The keyword response is {response.json()}')
    try:
      if response.json()['code']:
        created_kw_id = response.json()['data']['term_id']
      else: 
        created_kw_id = response.json()['id']
        print(f'The id is {created_kw_id}')
      return created_kw_id
    except TypeError:
      created_kw_id = ''
      return created_kw_id

原因解析 #

1. 异常类型不匹配：当访问字典中不存在的键时，Python抛出的是KeyError，但你的except块只指定了捕获TypeError，这个异常不在捕获范围内，所以会直接向上传播，最终导致Flask应用崩溃。
2. 重复解析响应：代码中多次调用response.json()，会重复解析响应体，虽然不会直接导致崩溃，但属于不必要的性能损耗，建议先将解析结果存入变量复用。

修复方案 #

方案1：扩展捕获的异常类型

将KeyError加入except块，确保能捕获到键不存在的情况：

def create_hec_kw(self, kw):
    print(f'Creating Keyword {kw}')
    data = {'name': kw, 'slug': kw.lower().replace(' ', '-')}
    response = requests.post(self.create_url('tags'), headers=self.get_headers(), json=data)
    response_json = response.json()  # 提前解析并存入变量
    print(f'The keyword response is {response_json}')
    try:
      if response_json['code']:
        created_kw_id = response_json['data']['term_id']
      else: 
        created_kw_id = response_json['id']
        print(f'The id is {created_kw_id}')
      return created_kw_id
    except (TypeError, KeyError):  # 同时捕获两种异常
      created_kw_id = ''
      return created_kw_id

方案2：使用字典get()方法避免异常

更稳妥的方式是用dict.get()方法访问键，不存在时返回None或指定默认值，从根源避免KeyError：

def create_hec_kw(self, kw):
    print(f'Creating Keyword {kw}')
    data = {'name': kw, 'slug': kw.lower().replace(' ', '-')}
    response = requests.post(self.create_url('tags'), headers=self.get_headers(), json=data)
    response_json = response.json()
    print(f'The keyword response is {response_json}')
    
    code = response_json.get('code')
    if code is not None:
        created_kw_id = response_json.get('data', {}).get('term_id', '')
    else:
        created_kw_id = response_json.get('id', '')
    
    if created_kw_id:
        print(f'The id is {created_kw_id}')
    return created_kw_id

内容的提问来源于stack exchange，提问作者user3324136