fix: propagate managed versioning for namespace connection (#3111)

Without this fix, if user directly use the native table to do operations
like `add_columns`, even if it is configured to use namespace db
connection, it is not really propagated through.

The fix is to bring lancedb's python binding up to date and do a similar
implementation as https://github.com/lance-format/lance/pull/5968, and
make sure the namespace is fully propagated through all the related
calls.

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Jack Ye
2026-03-09 01:44:31 -07:00
committed by GitHub
parent e93bb3355a
commit e0c5ceac03
17 changed files with 1042 additions and 125 deletions

View File

@@ -8,7 +8,7 @@ from abc import abstractmethod
from datetime import timedelta
from pathlib import Path
import sys
from typing import TYPE_CHECKING, Dict, Iterable, List, Literal, Optional, Union
from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Literal, Optional, Union
if sys.version_info >= (3, 12):
from typing import override
@@ -1541,6 +1541,8 @@ class AsyncConnection(object):
storage_options_provider: Optional["StorageOptionsProvider"] = None,
index_cache_size: Optional[int] = None,
location: Optional[str] = None,
namespace_client: Optional[Any] = None,
managed_versioning: Optional[bool] = None,
) -> AsyncTable:
"""Open a Lance Table in the database.
@@ -1573,6 +1575,9 @@ class AsyncConnection(object):
The explicit location (URI) of the table. If provided, the table will be
opened from this location instead of deriving it from the database URI
and table name.
managed_versioning: bool, optional
Whether managed versioning is enabled for this table. If provided,
avoids a redundant describe_table call when namespace_client is set.
Returns
-------
@@ -1587,6 +1592,8 @@ class AsyncConnection(object):
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
location=location,
namespace_client=namespace_client,
managed_versioning=managed_versioning,
)
return AsyncTable(table)

View File

@@ -12,7 +12,7 @@ from __future__ import annotations
import asyncio
import sys
from typing import Dict, Iterable, List, Optional, Union
from typing import Any, Dict, Iterable, List, Optional, Union
if sys.version_info >= (3, 12):
from typing import override
@@ -240,7 +240,7 @@ class LanceNamespaceDBConnection(DBConnection):
session : Optional[Session]
A session to use for this connection
"""
self._ns = namespace
self._namespace_client = namespace
self.read_consistency_interval = read_consistency_interval
self.storage_options = storage_options or {}
self.session = session
@@ -269,7 +269,7 @@ class LanceNamespaceDBConnection(DBConnection):
if namespace is None:
namespace = []
request = ListTablesRequest(id=namespace, page_token=page_token, limit=limit)
response = self._ns.list_tables(request)
response = self._namespace_client.list_tables(request)
return response.tables if response.tables else []
@override
@@ -309,7 +309,9 @@ class LanceNamespaceDBConnection(DBConnection):
# Try to describe the table first to see if it exists
try:
describe_request = DescribeTableRequest(id=table_id)
describe_response = self._ns.describe_table(describe_request)
describe_response = self._namespace_client.describe_table(
describe_request
)
location = describe_response.location
namespace_storage_options = describe_response.storage_options
except Exception:
@@ -323,7 +325,7 @@ class LanceNamespaceDBConnection(DBConnection):
location=None,
properties=self.storage_options if self.storage_options else None,
)
declare_response = self._ns.declare_table(declare_request)
declare_response = self._namespace_client.declare_table(declare_request)
if not declare_response.location:
raise ValueError(
@@ -353,7 +355,7 @@ class LanceNamespaceDBConnection(DBConnection):
# Only create if namespace returned storage_options (not None)
if storage_options_provider is None and namespace_storage_options is not None:
storage_options_provider = LanceNamespaceStorageOptionsProvider(
namespace=self._ns,
namespace=self._namespace_client,
table_id=table_id,
)
@@ -371,6 +373,7 @@ class LanceNamespaceDBConnection(DBConnection):
storage_options=merged_storage_options,
storage_options_provider=storage_options_provider,
location=location,
namespace_client=self._namespace_client,
)
return tbl
@@ -389,7 +392,7 @@ class LanceNamespaceDBConnection(DBConnection):
namespace = []
table_id = namespace + [name]
request = DescribeTableRequest(id=table_id)
response = self._ns.describe_table(request)
response = self._namespace_client.describe_table(request)
# Merge storage options: self.storage_options < user options < namespace options
merged_storage_options = dict(self.storage_options)
@@ -402,10 +405,14 @@ class LanceNamespaceDBConnection(DBConnection):
# Only create if namespace returned storage_options (not None)
if storage_options_provider is None and response.storage_options is not None:
storage_options_provider = LanceNamespaceStorageOptionsProvider(
namespace=self._ns,
namespace=self._namespace_client,
table_id=table_id,
)
# Pass managed_versioning to avoid redundant describe_table call in Rust.
# Convert None to False since we already have the answer from describe_table.
managed_versioning = response.managed_versioning is True
return self._lance_table_from_uri(
name,
response.location,
@@ -413,6 +420,8 @@ class LanceNamespaceDBConnection(DBConnection):
storage_options=merged_storage_options,
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
namespace_client=self._namespace_client,
managed_versioning=managed_versioning,
)
@override
@@ -422,7 +431,7 @@ class LanceNamespaceDBConnection(DBConnection):
namespace = []
table_id = namespace + [name]
request = DropTableRequest(id=table_id)
self._ns.drop_table(request)
self._namespace_client.drop_table(request)
@override
def rename_table(
@@ -484,7 +493,7 @@ class LanceNamespaceDBConnection(DBConnection):
request = ListNamespacesRequest(
id=namespace, page_token=page_token, limit=limit
)
response = self._ns.list_namespaces(request)
response = self._namespace_client.list_namespaces(request)
return ListNamespacesResponse(
namespaces=response.namespaces if response.namespaces else [],
page_token=response.page_token,
@@ -520,7 +529,7 @@ class LanceNamespaceDBConnection(DBConnection):
mode=_normalize_create_namespace_mode(mode),
properties=properties,
)
response = self._ns.create_namespace(request)
response = self._namespace_client.create_namespace(request)
return CreateNamespaceResponse(
properties=response.properties if hasattr(response, "properties") else None
)
@@ -555,7 +564,7 @@ class LanceNamespaceDBConnection(DBConnection):
mode=_normalize_drop_namespace_mode(mode),
behavior=_normalize_drop_namespace_behavior(behavior),
)
response = self._ns.drop_namespace(request)
response = self._namespace_client.drop_namespace(request)
return DropNamespaceResponse(
properties=(
response.properties if hasattr(response, "properties") else None
@@ -581,7 +590,7 @@ class LanceNamespaceDBConnection(DBConnection):
Response containing the namespace properties.
"""
request = DescribeNamespaceRequest(id=namespace)
response = self._ns.describe_namespace(request)
response = self._namespace_client.describe_namespace(request)
return DescribeNamespaceResponse(
properties=response.properties if hasattr(response, "properties") else None
)
@@ -615,7 +624,7 @@ class LanceNamespaceDBConnection(DBConnection):
if namespace is None:
namespace = []
request = ListTablesRequest(id=namespace, page_token=page_token, limit=limit)
response = self._ns.list_tables(request)
response = self._namespace_client.list_tables(request)
return ListTablesResponse(
tables=response.tables if response.tables else [],
page_token=response.page_token,
@@ -630,6 +639,8 @@ class LanceNamespaceDBConnection(DBConnection):
storage_options: Optional[Dict[str, str]] = None,
storage_options_provider: Optional[StorageOptionsProvider] = None,
index_cache_size: Optional[int] = None,
namespace_client: Optional[Any] = None,
managed_versioning: Optional[bool] = None,
) -> LanceTable:
# Open a table directly from a URI using the location parameter
# Note: storage_options should already be merged by the caller
@@ -643,6 +654,8 @@ class LanceNamespaceDBConnection(DBConnection):
)
# Open the table using the temporary connection with the location parameter
# Pass namespace_client to enable managed versioning support
# Pass managed_versioning to avoid redundant describe_table call
return LanceTable.open(
temp_conn,
name,
@@ -651,6 +664,8 @@ class LanceNamespaceDBConnection(DBConnection):
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
location=table_uri,
namespace_client=namespace_client,
managed_versioning=managed_versioning,
)
@@ -685,7 +700,7 @@ class AsyncLanceNamespaceDBConnection:
session : Optional[Session]
A session to use for this connection
"""
self._ns = namespace
self._namespace_client = namespace
self.read_consistency_interval = read_consistency_interval
self.storage_options = storage_options or {}
self.session = session
@@ -713,7 +728,7 @@ class AsyncLanceNamespaceDBConnection:
if namespace is None:
namespace = []
request = ListTablesRequest(id=namespace, page_token=page_token, limit=limit)
response = self._ns.list_tables(request)
response = self._namespace_client.list_tables(request)
return response.tables if response.tables else []
async def create_table(
@@ -750,7 +765,9 @@ class AsyncLanceNamespaceDBConnection:
# Try to describe the table first to see if it exists
try:
describe_request = DescribeTableRequest(id=table_id)
describe_response = self._ns.describe_table(describe_request)
describe_response = self._namespace_client.describe_table(
describe_request
)
location = describe_response.location
namespace_storage_options = describe_response.storage_options
except Exception:
@@ -764,7 +781,7 @@ class AsyncLanceNamespaceDBConnection:
location=None,
properties=self.storage_options if self.storage_options else None,
)
declare_response = self._ns.declare_table(declare_request)
declare_response = self._namespace_client.declare_table(declare_request)
if not declare_response.location:
raise ValueError(
@@ -797,7 +814,7 @@ class AsyncLanceNamespaceDBConnection:
and namespace_storage_options is not None
):
provider = LanceNamespaceStorageOptionsProvider(
namespace=self._ns,
namespace=self._namespace_client,
table_id=table_id,
)
else:
@@ -817,6 +834,7 @@ class AsyncLanceNamespaceDBConnection:
storage_options=merged_storage_options,
storage_options_provider=provider,
location=location,
namespace_client=self._namespace_client,
)
lance_table = await asyncio.to_thread(_create_table)
@@ -837,7 +855,7 @@ class AsyncLanceNamespaceDBConnection:
namespace = []
table_id = namespace + [name]
request = DescribeTableRequest(id=table_id)
response = self._ns.describe_table(request)
response = self._namespace_client.describe_table(request)
# Merge storage options: self.storage_options < user options < namespace options
merged_storage_options = dict(self.storage_options)
@@ -849,10 +867,14 @@ class AsyncLanceNamespaceDBConnection:
# Create a storage options provider if not provided by user
if storage_options_provider is None and response.storage_options is not None:
storage_options_provider = LanceNamespaceStorageOptionsProvider(
namespace=self._ns,
namespace=self._namespace_client,
table_id=table_id,
)
# Capture managed_versioning from describe response.
# Convert None to False since we already have the answer from describe_table.
managed_versioning = response.managed_versioning is True
# Open table in a thread
def _open_table():
temp_conn = LanceDBConnection(
@@ -870,6 +892,8 @@ class AsyncLanceNamespaceDBConnection:
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
location=response.location,
namespace_client=self._namespace_client,
managed_versioning=managed_versioning,
)
lance_table = await asyncio.to_thread(_open_table)
@@ -881,7 +905,7 @@ class AsyncLanceNamespaceDBConnection:
namespace = []
table_id = namespace + [name]
request = DropTableRequest(id=table_id)
self._ns.drop_table(request)
self._namespace_client.drop_table(request)
async def rename_table(
self,
@@ -943,7 +967,7 @@ class AsyncLanceNamespaceDBConnection:
request = ListNamespacesRequest(
id=namespace, page_token=page_token, limit=limit
)
response = self._ns.list_namespaces(request)
response = self._namespace_client.list_namespaces(request)
return ListNamespacesResponse(
namespaces=response.namespaces if response.namespaces else [],
page_token=response.page_token,
@@ -978,7 +1002,7 @@ class AsyncLanceNamespaceDBConnection:
mode=_normalize_create_namespace_mode(mode),
properties=properties,
)
response = self._ns.create_namespace(request)
response = self._namespace_client.create_namespace(request)
return CreateNamespaceResponse(
properties=response.properties if hasattr(response, "properties") else None
)
@@ -1012,7 +1036,7 @@ class AsyncLanceNamespaceDBConnection:
mode=_normalize_drop_namespace_mode(mode),
behavior=_normalize_drop_namespace_behavior(behavior),
)
response = self._ns.drop_namespace(request)
response = self._namespace_client.drop_namespace(request)
return DropNamespaceResponse(
properties=(
response.properties if hasattr(response, "properties") else None
@@ -1039,7 +1063,7 @@ class AsyncLanceNamespaceDBConnection:
Response containing the namespace properties.
"""
request = DescribeNamespaceRequest(id=namespace)
response = self._ns.describe_namespace(request)
response = self._namespace_client.describe_namespace(request)
return DescribeNamespaceResponse(
properties=response.properties if hasattr(response, "properties") else None
)
@@ -1072,7 +1096,7 @@ class AsyncLanceNamespaceDBConnection:
if namespace is None:
namespace = []
request = ListTablesRequest(id=namespace, page_token=page_token, limit=limit)
response = self._ns.list_tables(request)
response = self._namespace_client.list_tables(request)
return ListTablesResponse(
tables=response.tables if response.tables else [],
page_token=response.page_token,

View File

@@ -1746,6 +1746,8 @@ class LanceTable(Table):
storage_options_provider: Optional["StorageOptionsProvider"] = None,
index_cache_size: Optional[int] = None,
location: Optional[str] = None,
namespace_client: Optional[Any] = None,
managed_versioning: Optional[bool] = None,
_async: AsyncTable = None,
):
if namespace is None:
@@ -1753,6 +1755,7 @@ class LanceTable(Table):
self._conn = connection
self._namespace = namespace
self._location = location # Store location for use in _dataset_path
self._namespace_client = namespace_client
if _async is not None:
self._table = _async
else:
@@ -1764,6 +1767,8 @@ class LanceTable(Table):
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
location=location,
namespace_client=namespace_client,
managed_versioning=managed_versioning,
)
)
@@ -1806,6 +1811,8 @@ class LanceTable(Table):
storage_options_provider: Optional["StorageOptionsProvider"] = None,
index_cache_size: Optional[int] = None,
location: Optional[str] = None,
namespace_client: Optional[Any] = None,
managed_versioning: Optional[bool] = None,
):
if namespace is None:
namespace = []
@@ -1817,6 +1824,8 @@ class LanceTable(Table):
storage_options_provider=storage_options_provider,
index_cache_size=index_cache_size,
location=location,
namespace_client=namespace_client,
managed_versioning=managed_versioning,
)
# check the dataset exists
@@ -1848,6 +1857,16 @@ class LanceTable(Table):
"Please install with `pip install pylance`."
)
if self._namespace_client is not None:
table_id = self._namespace + [self.name]
return lance.dataset(
version=self.version,
storage_options=self._conn.storage_options,
namespace=self._namespace_client,
table_id=table_id,
**kwargs,
)
return lance.dataset(
self._dataset_path,
version=self.version,
@@ -2713,6 +2732,7 @@ class LanceTable(Table):
data_storage_version: Optional[str] = None,
enable_v2_manifest_paths: Optional[bool] = None,
location: Optional[str] = None,
namespace_client: Optional[Any] = None,
):
"""
Create a new table.
@@ -2773,6 +2793,7 @@ class LanceTable(Table):
self._conn = db
self._namespace = namespace
self._location = location
self._namespace_client = namespace_client
if data_storage_version is not None:
warnings.warn(