Python Kafka消费者读取已读消息

Ish*_*ema 2 python apache-kafka kafka-consumer-api

卡夫卡消费者代码 -

def test():
TOPIC = "file_data"
producer = KafkaProducer()
producer.send(TOPIC, "data")
consumer = KafkaConsumer(
    bootstrap_servers=['localhost:9092'],
    auto_offset_reset='latest',
    consumer_timeout_ms=1000,
    group_id="Group2",
    enable_auto_commit=False,
    auto_commit_interval_ms=1000
)
topic_partition = TopicPartition(TOPIC, 0)
assigned_topic = [topic_partition]
consumer.assign(assigned_topic)
consumer.seek_to_beginning(topic_partition)
for message in consumer:
    print("%s key=%s value=%s" % (message.topic, message.key, message.value))
consumer.commit()
Run Code Online (Sandbox Code Playgroud)

预期行为 它应该只读取生产者写入的最后一条消息。它应该只打印:

file_data key=None value=b'data'
Run Code Online (Sandbox Code Playgroud)

当前行为 运行代码后打印:

file_data key=None value=b'data'
file_data key=None value=b'data'
file_data key=None value=b'data'
file_data key=None value=b'data'
file_data key=None value=b'data'
file_data key=None value=b'data'
Run Code Online (Sandbox Code Playgroud)

小智 5

from kafka import KafkaConsumer
from kafka import TopicPartition
from kafka import KafkaProducer

def test():
    TOPIC = "file_data"
    producer = KafkaProducer()
    producer.send(TOPIC, b'data')
    consumer = KafkaConsumer(
        bootstrap_servers=['localhost:9092'],
        auto_offset_reset='latest',
        consumer_timeout_ms=1000,
        group_id="Group2",
        enable_auto_commit=False,
        auto_commit_interval_ms=1000
    )
    topic_partition = TopicPartition(TOPIC, 0)
    assigned_topic = [topic_partition]
    consumer.assign(assigned_topic)
    # consumer.seek_to_beginning(topic_partition)
    for message in consumer:
        print("%s key=%s value=%s" % (message.topic, message.key, message.value))
    consumer.commit()
test()

Run Code Online (Sandbox Code Playgroud)

这正在按照您的期望进行。如果你想让它从头开始,那么只需调用seekToBeginning

参考:seek_to_beginning