Clarification Needed on NewTopic Usage with Custom Replica Assignment in Confluent Kafka Python API

[AkshatHyperion]

Hello,

I am using the Confluent Kafka Python API to create topics with custom replica assignments and have encountered some confusion regarding the usage of NewTopic, specifically when setting num_partitions and replication_factor in conjunction with replica_assignment.

According to the documentation, when using replica_assignment, both num_partitions and replication_factor should be set to -1. However, when I do this, I encounter an error:

Failed to create topic: Specifying a default partition count and a replica assignment are mutually exclusive

from confluent_kafka.admin import AdminClient, NewTopic

config = {'bootstrap.servers': 'your_broker_address'}
admin_client = AdminClient(config)

topic_name = 'temp'
replica_assignment = [[1]]  # Example: One partition with its replica on broker 1

new_topic = NewTopic(topic_name, num_partitions=-1, replication_factor=-1, replica_assignment=replica_assignment)

try:
    fs = admin_client.create_topics([new_topic])
    # Handling the result...
except Exception as e:
    print(f"Failed to create topic: {e}")

Interestingly, when I set num_partitions to 1 instead of -1, the topic creation succeeds without any errors:

new_topic = NewTopic(topic_name, num_partitions=1, replication_factor=-1, replica_assignment=replica_assignment)

Could you please clarify the correct usage in this scenario? Is the documentation accurate, or is there an update or a specific context where num_partitions should not be -1 when using replica_assignment?

Thank you for your help!