本地运行AWS Lambda函数报string indices must be integers错误

错误复现信息 #

本地执行Lambda函数时触发如下类型错误：

String indices must be integers

错误栈定位异常触发位置为代码行payload = loads(record['body'], parse_float=str)。

对应Lambda函数代码 #

def lambda_handler(event, context):

    logger.debug('Received event: {}'.format(event))
    print(event['Item']) 
    operations = {
        'DELETE': lambda dynamo, x: dynamo.delete_item(**x),
        'POST': lambda dynamo, x: dynamo.put_item(**x),
        'PUT': lambda dynamo, x: dynamo.update_item(**x),
        'GET': lambda dynamo, x: dynamo.get_item(**x),
        'GET_ALL': lambda dynamo, x: dynamo.scan(**x),
        'BATCH_WRITE': lambda dynamo, x: dynamo.batch_write_item(**x),
    }
    for record in event['Item']:
        print(record['body'])
        payload = loads(record['body'], parse_float=str)
        operation = record['messageAttributes']['Method']['stringValue']
        if operation in operations:
            try:
                operations[operation](dynamo_client, payload)
                logger.info('{} method successful'.format(operation))
                logger.debug('Event payload: {}'.format(payload))
            except Exception as e:
                logger.error(e)
        else:
            logger.error('Unsupported method \'{}\''.format(operation))
    return 0

本地测试传入的事件JSON #

{ 
    "TableName":"DemoTable",
    "Item":
            {"make":{"S":"Ford"},
             "type":{"S":"Sedan"},
             "fuelType":{"S":"gas"},
             "miles":{"S":"50000"},
             "model":{"S":"Fiesta"},
             "year":{"S":"2018"}}
}

完整错误日志 #

[DEBUG] 2022-07-03T15:09:24.716Z         Received event: {'TableName': 'DemoTable', 'Item': {'make': {'S': 'Ford'}, 'type': {'S': 'Sedan'}, 'fuelType': {'S': 'gas'}, 'miles': {'S': '50000'}, 'model': {'S': 'Fiesta'}, 'year': {'S': '2018'}}}
{'make': {'S': 'Ford'}, 'type': {'S': 'Sedan'}, 'fuelType': {'S': 'gas'}, 'miles': {'S': '50000'}, 'model': {'S': 'Fiesta'}, 'year': {'S': '2018'}}
payload = loads(record['body'], parse_float=str)r
END
{"errorMessage": "string indices must be integers", "errorType": "TypeError",  "stackTrace": ["  File \"/var/task/app.py\", line 29, in lambda_handler\n    payload = loads(record['body'], parse_float=str)\n"]}

错误根因 #

• 直接触发报错的语法问题：Python中直接对字典做for x in dict迭代时，循环拿到的x是字典的键，不是键对应的值，也不是键值对元组。传入的event['Item']是存储车辆属性的字典，迭代时依次返回的是'make'、'type'、'fuelType'这类字符串值。后续代码对字符串类型的record执行record['body']操作时，Python规定字符串的索引必须是整数（比如取字符串第n位用s[n]），传入字符串'body'作为索引就会直接抛出String indices must be integers错误。
• 深层逻辑不匹配问题：现有循环处理逻辑是适配SQS触发Lambda的事件格式的——这种场景下事件里的记录列表应该是数组类型，每个数组元素是单条消息，包含body（序列化的消息内容）、messageAttributes（消息自定义属性）字段。但本地传入的Item是DynamoDB单条数据格式的字典，和代码预期的结构完全不匹配，就算修复了迭代的语法问题，后续也会因为找不到body、messageAttributes字段继续报错。

修复方案 #

根据实际业务场景二选一即可：

• 场景1：Lambda仅需处理单条DynamoDB数据写入，不需要支持多操作、批量消息
  删掉适配SQS批量消息的循环逻辑，直接从event中提取参数构造DynamoDB请求即可，参考代码：

  from json import loads
  import boto3
  import logging
  logger = logging.getLogger()
  logger.setLevel(logging.INFO)
  dynamo_client = boto3.client('dynamodb')
  
  def lambda_handler(event, context):
      logger.debug('Received event: {}'.format(event))
      payload = {
          "TableName": event["TableName"],
          "Item": event["Item"]
      }
      try:
          dynamo_client.put_item(**payload)
          logger.info('POST method successful')
          logger.debug('Event payload: {}'.format(payload))
      except Exception as e:
          logger.error(e)
          return -1
      return 0
  

• 场景2：需要保留原有支持多DynamoDB操作、批量处理消息的逻辑
  把本地测试的事件结构改成代码预期的格式：Item字段为数组类型，每个数组元素包含序列化后操作参数的body字段、标记操作类型的messageAttributes字段，参考测试事件：

  {
      "Item": [
          {
              "body": "{\"TableName\": \"DemoTable\", \"Item\": {\"make\": {\"S\": \"Ford\"}, \"type\": {\"S\": \"Sedan\"}, \"fuelType\": {\"S\": \"gas\"}, \"miles\": {\"S\": \"50000\"}, \"model\": {\"S\": \"Fiesta\"}, \"year\": {\"S\": \"2018\"}}}",
              "messageAttributes": {
                  "Method": {
                      "stringValue": "POST"
                  }
              }
          }
      ]
  }
  

内容的提问来源于stack exchange，提问作者Sam