無法使用Kafka-Python的反序列化器從Kafka消耗JSON消息

Question

我試圖通過Kafka發送一個非常簡單的JSON對象，並使用Python和kafka-python將其讀出來。 但是，我一直看到以下錯誤：

2017-04-07 10:28:52,030.30.9998989105:kafka.future:8228:ERROR:10620:Error processing callback
Traceback (most recent call last):
  File "C:\Anaconda2\lib\site-packages\kafka\future.py", line 79, in _call_backs
    f(value)
  File "C:\Anaconda2\lib\site-packages\kafka\consumer\fetcher.py", line 760, in _handle_fetch_response
    unpacked = list(self._unpack_message_set(tp, messages))
  File "C:\Anaconda2\lib\site-packages\kafka\consumer\fetcher.py", line 539, in _unpack_message_set
    tp.topic, msg.value)
  File "C:\Anaconda2\lib\site-packages\kafka\consumer\fetcher.py", line 570, in _deserialize
    return f(bytes_)
  File "C:\Users\myUser\workspace\PythonKafkaTest\src\example.py", line 55, in <lambda>
    value_deserializer=lambda m: json.loads(m).decode('utf-8'))
  File "C:\Anaconda2\lib\json\__init__.py", line 339, in loads
    return _default_decoder.decode(s)
  File "C:\Anaconda2\lib\json\decoder.py", line 364, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "C:\Anaconda2\lib\json\decoder.py", line 382, in raw_decode
    raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded

我做了一些研究，這個錯誤最常見的原因是JSON錯了。 我在發送之前嘗試打印出JSON，方法是將以下內容添加到我的代碼中，並且JSON打印沒有錯誤。

  while True:
        json_obj1 = json.dumps({"dataObjectID": "test1"})
        print json_obj1
        producer.send('my-topic', {"dataObjectID": "test1"})
        producer.send('my-topic', {"dataObjectID": "test2"})
        time.sleep(1)

這讓我懷疑我可以生產json，但不會消耗它。

這是我的代碼：

import threading
import logging
import time
import json

from kafka import KafkaConsumer, KafkaProducer


class Producer(threading.Thread):
    daemon = True

    def run(self):
        producer = KafkaProducer(bootstrap_servers='localhost:9092',
                                 value_serializer=lambda v: json.dumps(v).encode('utf-8'))

        while True:
            producer.send('my-topic', {"dataObjectID": "test1"})
            producer.send('my-topic', {"dataObjectID": "test2"})
            time.sleep(1)


class Consumer(threading.Thread):
    daemon = True

    def run(self):
        consumer = KafkaConsumer(bootstrap_servers='localhost:9092',
                                 auto_offset_reset='earliest',
                                 value_deserializer=lambda m: json.loads(m).decode('utf-8'))
        consumer.subscribe(['my-topic'])

        for message in consumer:
            print (message)


def main():
    threads = [
        Producer(),
        Consumer()
    ]

    for t in threads:
        t.start()

    time.sleep(10)

if __name__ == "__main__":
    logging.basicConfig(
        format='%(asctime)s.%(msecs)s:%(name)s:%(thread)d:' +
               '%(levelname)s:%(process)d:%(message)s',
        level=logging.INFO
    )
    main()

如果刪除value_serializer和value_deserializer，我可以成功發送和接收字符串。 當我運行該代碼時，我可以看到我發送的JSON。這是一個簡短的片段：

ConsumerRecord(topic=u'my-topic', partition=0, offset=5742, timestamp=None, timestamp_type=None, key=None, value='{"dataObjectID": "test1"}', checksum=-1301891455, serialized_key_size=-1, serialized_value_size=25)
ConsumerRecord(topic=u'my-topic', partition=0, offset=5743, timestamp=None, timestamp_type=None, key=None, value='{"dataObjectID": "test2"}', checksum=-1340077864, serialized_key_size=-1, serialized_value_size=25)
ConsumerRecord(topic=u'my-topic', partition=0, offset=5744, timestamp=None, timestamp_type=None, key=None, value='test', checksum=1495943047, serialized_key_size=-1, serialized_value_size=4)
ConsumerRecord(topic=u'my-topic', partition=0, offset=5745, timestamp=None, timestamp_type=None, key=None, value='\xc2Hello, stranger!', checksum=-1090450220, serialized_key_size=-1, serialized_value_size=17)
ConsumerRecord(topic=u'my-topic', partition=0, offset=5746, timestamp=None, timestamp_type=None, key=None, value='test', checksum=1495943047, serialized_key_size=-1, serialized_value_size=4)
ConsumerRecord(topic=u'my-topic', partition=0, offset=5747, timestamp=None, timestamp_type=None, key=None, value='\xc2Hello, stranger!', checksum=-1090450220, serialized_key_size=-1, serialized_value_size=17)

所以我嘗試從使用者中刪除value_deserializer，並且該代碼執行但沒有反序列化器，消息以String形式出現，這不是我需要的。 那么，為什么value_deserializer不起作用呢？ 是否有不同的方法從我應該使用的Kafka消息中獲取JSON？

Answer 1

首先將消息解碼為utf-8，然后json.load / dump它解決了我的問題：

value_deserializer=lambda m: json.loads(m.decode('utf-8'))

代替：

value_deserializer=lambda m: json.loads(m).decode('utf-8')

希望這對制片人也有效

Answer 2

事實證明問題是value_deserializer=lambda m: json.loads(m).decode('utf-8')的解碼部分，當我將其更改為value_deserializer=lambda m: json.loads(m)然后我看到從Kafka讀取的對象類型現在是一個字典。 基於python的JSON文檔中的以下信息是正確的：

|---------------------|------------------|
|       JSON          |     Python       |
|---------------------|------------------|
|      object         |      dict        |
|---------------------|------------------|
|      array          |      list        |
|---------------------|------------------|
|      string         |      unicode     |
|---------------------|------------------|
|      number (int)   |      int, long   |
|---------------------|------------------|
|      number (real)  |      float       |
|---------------------|------------------|
|      true           |      True        |
|---------------------|------------------|
|      false          |      False       |
|---------------------|------------------|
|      null           |      None        |
|---------------------|------------------|

Answer 3

你不需要lambda ...而不是

value_deserializer=lambda m: json.loads(m)

你應該使用

value_deserializer=json.load