python 读写kafka

MusicDancing

已于 2024-01-16 11:31:54 修改

阅读量707

点赞数 9

CC 4.0 BY-SA版权

分类专栏： python 文章标签： python kafka 开发语言

于 2024-01-15 20:07:15 首次发布

本文链接：https://blog.csdn.net/MusicDancing/article/details/135609740

python 专栏收录该内容

28 篇文章

订阅专栏

本文详细介绍了如何在Python中使用Pykafka库进行Kafka的消息生产和消费，包括安装步骤、生产者代码实例和消费者代码实例，以及相关的参数配置。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

1. 安装pykafka

pip install pykafka

2. 生产者

from pykafka import KafkaClient
 
def get_kafka_producer(hosts, topics):
    client = KafkaClient(hosts=hosts)
    print(client.topics)
    topic = client.topics[topics]
    producer = topic.get_producer()
    return producer

测试

hosts = '192.168.20.203:9092,192.168.20.204:9092,192.168.20.205:9092'
topics = "test_kafka_topic"
producer = get_kafka_producer(hosts, topics)   
for i in range(10):
    msg = "test message " + str(i)
    # msg = bytes(msg, encoding='utf-8')    
    # producer.produce(msg)
    producer.produce(msg.encode())
producer.stop()

3. 消费者

def get_kafka_consumer(hosts, topics):
    client = KafkaClient(hosts=hosts)
    topic=client.topics[topics]
    consumer = topic.get_balanced_consumer(consumer_group='test_kafka_topic', auto_commit_enable=True,
zookeeper_connect='192.168.20.201:2181,192.168.20.202:2181,192.168.20.203:2181',
managed=True, consumer_timeout_ms=1000)
    # managed=True，即使用新式reblance分区方法，不需zk；managed=False则需通过zk来实现reblance      
    return consumer

测试

hosts = '192.168.20.203:9092,192.168.20.204:9092,192.168.20.205:9092'
topics = "test_kafka_topic"
consumer = get_kafka_consumer(hosts, topics)   
for msg in consumer:
    print(msg)
    if msg is not None:
        print(msg.offset)
        print(msg.value)