Queue Management¶

Creating a Queue¶

queue.create_queue("my_queue")

Unlogged Queues¶

Unlogged queues are faster but do not survive a database crash. Use them for transient, high-throughput data.

queue.create_queue("temp_queue", unlogged=True)

Partitioned Queues¶

Partitioned queues split data across multiple tables, improving performance for very large queues.

queue.create_partitioned_queue(
    "big_queue",
    partition_interval=10000,
    retention_interval=100000,
)

Parameter	Description
`partition_interval`	Number of messages per partition.
`retention_interval`	Number of messages to retain in the active partition set.

Dropping a Queue¶

dropped: bool = queue.drop_queue("my_queue")

Dropping a queue removes the queue table and its archive table. Returns True on success.

Listing Queues¶

queues = queue.list_queues()
for q in queues:
    print(q.queue_name, q.is_partitioned, q.created_at)

list_queues() returns a list of QueueRecord objects:

Attribute	Type	Description
`queue_name`	`str`	Name of the queue.
`is_partitioned`	`bool`	Whether the queue is partitioned.
`is_unlogged`	`bool`	Whether the queue is unlogged.
`created_at`	`datetime`	Creation timestamp.

Backward compatibility note: Older versions returned List[str]. The current version returns List[QueueRecord] and emits a UserWarning on the sync client. Use .queue_name to get the string name.

Purging a Queue¶

Remove all messages from a queue without deleting the queue itself:

purged_count: int = queue.purge("my_queue")

Validating a Queue Name¶

queue.validate_queue_name("my_queue")  # Raises if invalid

This delegates validation to pgmq.validate_queue_name() in the database, ensuring your queue name complies with PGMQ naming rules.