compact()

This operation starts a compaction job that merges small segments in a collection to improve storage layout and query efficiency.

Request Syntax

compact(
    collection_name: str,
    is_clustering: Optional[bool] = False,
    is_l0: Optional[bool] = False,
    target_size: Optional[int] = None,
    target_size_unit: str = "mb",
    timeout: Optional[float] = None,
    **kwargs,
) -> int

PARAMETERS:

• collection_name (str) -

  [REQUIRED]

  The name of the target collection.

• is_clustering (bool) -

  Whether to perform a clustering compaction. Defaults to False.

• is_l0 (bool) -

  Whether to perform an L0 compaction, which specifically handles L0 segments by merging delete operations into existing data segments. Defaults to False.

• target_size (int) -

  Target segment size after compaction. Must be a positive integer. The unit is specified by target_size_unit (default MB). If not provided, the server uses its default target size.

• target_size_unit (str) -

Unit for target_size. Supported values: "b", "kb", “mb” (default), "gb", "tb", "pb". The value is converted to MB before being sent to the server.

• timeout (float) -

  The timeout duration for this operation. Setting this to None indicates that this operation timeouts when any response arrives or any error occurs.

• kwargs (dict) -

  Optional request context parameters.

RETURN TYPE:

int

RETURNS:

Compaction job ID for follow-up status queries.

A compaction job ID, which can be used to get the compaction status.

EXCEPTIONS:

• ParamError

  Raised when target_size is not an integer or when target_size_unit is invalid.

• MilvusException

  This exception will be raised when any error occurs during this operation.

Examples

from pymilvus import MilvusClient

client = MilvusClient(uri="http://localhost:19530", token="root:Milvus")

# Create a collection and insert some data
client.create_collection(collection_name="book_catalog", dimension=4)
client.insert(collection_name="book_catalog", data=[{"id": i, "vector": [0.1] * 4} for i in range(10)])

# Flush the data so that the inserted segments are sealed and become compactable
client.flush(collection_name="book_catalog")

job_id = client.compact(
    collection_name="book_catalog",
    target_size=1024,
    target_size_unit="mb",
)

print(job_id)