Kafka-python 获取主题的分区数

Question

I'm using: https://github.com/mumrah/kafka-python as a kafka api in Python. I want to get the number of partitions for a specified topic. How do I do that?

Answer 1

Might be a slightly simplistic solution, but:

from kafka import KafkaClient

client = KafkaClient('SERVER:PORT')
topic_partition_ids = client.get_partition_ids_for_topic(b'TOPIC')
len(topic_partition_ids)

tested on Python 3.4.3 / kafka-python 0.9.3

Answer 2

I found this question while trying to solve this exact same problem. I know the question is old, but here is the solution I came up with (using Kazoo to talk to zookeeper):

from kazoo.client import KazooClient

class KafkaInfo(object):
    def __init__(self, hosts):
        self.zk = KazooClient(hosts)
        self.zk.start()

    def topics(self):
        return self.zk.get_children('/brokers/topics')

    def partitions(self, topic):
        strs = self.zk.get_children('/brokers/topics/%s/partitions' % topic)
        return map(int, strs)

    def consumers(self):
        return self.zk.get_children('/consumers')

    def topics_for_consumer(self, consumer):
        return self.zk.get_children('/consumers/%s/offsets' % consumer)

    def offset(self, topic, consumer, partition):
        (n, _) = self.zk.get('/consumers/%s/offsets/%s/%d' % (consumer, topic, partition))
        return int(n)

Answer 3

Python 3.8.10/kafka-python 2.0.2 Solution:

from kafka import KafkaConsumer

def get_partitions_number(server, topic):
    consumer = KafkaConsumer(
        topic,
        bootstrap_servers=server
    )
    partitions = consumer.partitions_for_topic(topic)
    return len(partitions)

partitions_for_topic

Answer 4

For those of you using Confluent-Python or the enterprise API. This can be done this way:

def count_partitions(my_partitions) -> int:
    count = 0
    for part in my_partitions:
        count = count + 1
    return count

cluster_data: ClusterMetadata = producer.list_topics(topic=TOPIC)
    topic_data: TopicMetadata = cluster_data.topics[TOPIC]
    available_partitions: PartitionMetadata = topic_data.partitions
    print(count_partitions(available_partitions))