fix(kernel): preserve empty metadata filters

CHANGELOG.md

@@ -1,6 +1,7 @@
 # Release History
 
 # Unreleased
+- Kernel metadata filters are now forwarded unchanged instead of collapsing empty strings to `None`. Only `None` leaves a filter unset; empty pattern filters match nothing (PECOBLR-4221).
 - Kernel backend (`use_kernel=True`): OAuth **M2M with a JWT private-key client assertion** (RFC 7523) is now supported. Pass `oauth_client_id` + `oauth_jwt_key_file` + `oauth_jwt_kid` (with optional `oauth_jwt_passphrase` for an encrypted PKCS#8 key, `oauth_jwt_algorithm` defaulting to `RS256`, `oauth_scopes`, and `token_url` for the IdP token endpoint) and the connector routes them to the kernel's `auth_type="oauth-m2m-jwt"`, which signs a short-lived assertion with the private key instead of sending a client secret. The kernel owns the token lifecycle. A private-key file is treated as unambiguous JWT M2M intent and is mutually exclusive with `oauth_client_secret` / `credentials_provider` (both raise `NotSupportedError`). Verified end-to-end against an Azure Databricks workspace with the service principal's public certificate registered on its Entra ID app registration. Requires `databricks-sql-kernel >= 0.2.0` with JWT support.
 - Kernel backend (`use_kernel=True`): OAuth U2M with `auth_type="databricks-oauth"` now forwards the connector's `databricks-sql-python` OAuth-app bundle (`client_id` + `sql offline_access` scopes + redirect port) into the kernel, so a bare U2M connection authenticates as `databricks-sql-python` — parity with the Thrift path — instead of inheriting the kernel's own `databricks-sql-connector` default. A caller-supplied `oauth_client_id` (with its coupled `oauth_redirect_port`) is honored, as is a caller-supplied `oauth_scopes`; absent one, the connector default (`sql offline_access`) is forwarded. Note: the kernel binds a single U2M redirect port, so unlike the Thrift path (which tries the full `8020..8024` range) the kernel path uses only one port and does not fall back to the next port if it is already bound — pass `oauth_redirect_port` (with `oauth_client_id`) to pick a free one on a port collision (PECOBLR-4040)
 - Kernel backend (`use_kernel=True`): **Azure Entra (Azure AD) service-principal M2M is now supported.** `auth_type="azure-sp-m2m"` forwards `azure_client_id` / `azure_client_secret`; the kernel is the Azure-aware auth core — it builds the Entra v2.0 token endpoint and the `{app_id}/.default` scope, and **auto-discovers the tenant** from the workspace's `/aad/auth` redirect when `azure_tenant_id` is omitted (matching Thrift). The `Authorization` bearer is the Databricks-audience data token, which alone authenticates a workspace-member SP. Set `azure_workspace_resource_id` and the kernel also sends the Azure SP management token (`X-Databricks-Azure-SP-Management-Token`) + `X-Databricks-Azure-Workspace-Resource-Id` header (matching the JDBC driver), so a service principal with an Azure RBAC role but no workspace membership can authenticate; omit it and no ARM management-scope token is fetched. Azure AD **U2M** (`auth_type="azure-oauth"`) now routes to the kernel's OAuth U2M flow, identically to `auth_type="databricks-oauth"`: the kernel runs the in-house workspace-federated browser flow, which Azure workspaces support (the workspace federates login to Entra). It forwards the connector's `databricks-sql-python` OAuth app, not the Thrift Azure app (`96eecda7` / port 8030), which is registered for Thrift's direct-Entra flow the kernel does not perform (PECOBLR-4141; PECOBLR-4120)

src/databricks/sql/backend/databricks_client.py

@@ -248,8 +248,10 @@ def get_schemas(
             max_rows: Maximum number of rows to fetch in a single batch
             max_bytes: Maximum number of bytes to fetch in a single batch
             cursor: The cursor object that will handle the results
-            catalog_name: Optional catalog name pattern to filter by
-            schema_name: Optional schema name pattern to filter by
+            catalog_name: Optional exact catalog name to filter by, forwarded
+                unchanged. ``None`` leaves the filter unset.
+            schema_name: Optional schema name pattern to filter by. ``None``
+                leaves the filter unset; an empty string matches nothing.
 
         Returns:
             ResultSet: An object containing the schema metadata
@@ -284,10 +286,13 @@ def get_tables(
             max_bytes: Maximum number of bytes to fetch in a single batch
             cursor: The cursor object that will handle the results
             catalog_name: Optional catalog name pattern to filter by
-                if catalog_name is None, we fetch across all catalogs
+                if catalog_name is None, we fetch across all catalogs; an empty
+                string matches nothing
             schema_name: Optional schema name pattern to filter by
-                if schema_name is None, we fetch across all schemas
-            table_name: Optional table name pattern to filter by
+                if schema_name is None, we fetch across all schemas; an empty
+                string matches nothing
+            table_name: Optional table name pattern to filter by. ``None``
+                leaves the filter unset; an empty string matches nothing.
             table_types: Optional list of table types to filter by (e.g., ['TABLE', 'VIEW'])
 
         Returns:
@@ -322,11 +327,15 @@ def get_columns(
             max_rows: Maximum number of rows to fetch in a single batch
             max_bytes: Maximum number of bytes to fetch in a single batch
             cursor: The cursor object that will handle the results
-            catalog_name: Optional catalog name pattern to filter by
-            schema_name: Optional schema name pattern to filter by
+            catalog_name: Optional exact catalog name to filter by, forwarded
+                unchanged. ``None`` leaves the filter unset.
+            schema_name: Optional schema name pattern to filter by. ``None``
+                leaves the filter unset; an empty string matches nothing.
             table_name: Optional table name pattern to filter by
-                if table_name is None, we fetch across all tables
-            column_name: Optional column name pattern to filter by
+                if table_name is None, we fetch across all tables; an empty
+                string matches nothing
+            column_name: Optional column name pattern to filter by. ``None``
+                leaves the filter unset; an empty string matches nothing.
 
         Returns:
             ResultSet: An object containing the column metadata

src/databricks/sql/backend/kernel/client.py

@@ -117,39 +117,6 @@ def _is_not_found(exc: BaseException) -> bool:
     )
 
 
-def _none_if_blank(value: Optional[str]) -> Optional[str]:
-    """Map an empty/whitespace-only metadata filter to ``None``
-    ("match all"), matching the Thrift backend's effective behaviour.
-
-    The kernel's ``Identifier`` / ``LikePattern`` reject ``""`` with
-    ``InvalidArgument`` (-> ``ProgrammingError``); ``None`` is the
-    kernel's canonical "match all". Applied to schema / table / column
-    *pattern* args (which otherwise keep ``%`` / ``_`` as real LIKE
-    wildcards)."""
-    if value is None:
-        return None
-    return value if value.strip() else None
-
-
-def _catalog_or_none(value: Optional[str]) -> Optional[str]:
-    """Normalise a catalog filter: ``None`` / blank / ``'%'`` / ``'*'``
-    all mean "all catalogs" -> ``None``.
-
-    This makes ``columns(catalog='%')`` behave like
-    ``tables(catalog='%')`` / ``schemas(catalog='%')`` — the kernel
-    already treats blank/``%``/``*`` as "all catalogs" for SHOW SCHEMAS
-    / SHOW TABLES (``is_null_or_wildcard``) but treats the catalog as an
-    exact identifier for SHOW COLUMNS, so the three diverged. Normalising
-    connector-side makes them symmetric. This intentionally diverges from
-    raw-Thrift literalness (Thrift treats ``%`` as a literal catalog
-    name) in favour of JDBC "catalog is exact-or-all, not a pattern" +
-    internal consistency. Catalog is the only arg normalised this way;
-    schema/table/column patterns keep ``%`` / ``*`` as LIKE wildcards."""
-    if value is None or not value.strip() or value in ("%", "*"):
-        return None
-    return value
-
-
 def _is_staging_statement(operation: str) -> bool:
     """True iff ``operation`` is a volume/staging statement (PUT / GET /
     REMOVE).
@@ -937,8 +904,8 @@ def get_schemas(
             raise InterfaceError("get_schemas requires an open session.")
         try:
             stream = self._kernel_session.metadata().list_schemas(
-                catalog=_catalog_or_none(catalog_name),
-                schema_pattern=_none_if_blank(schema_name),
+                catalog=catalog_name,
+                schema_pattern=schema_name,
             )
             return self._make_result_set(stream, cursor, self._synthetic_command_id())
         except Exception as exc:
@@ -964,9 +931,9 @@ def get_tables(
             # do the work — no connector-side drain + refilter. Passing it
             # through preserves streaming for large schemas.
             stream = self._kernel_session.metadata().list_tables(
-                catalog=_catalog_or_none(catalog_name),
-                schema_pattern=_none_if_blank(schema_name),
-                table_pattern=_none_if_blank(table_name),
+                catalog=catalog_name,
+                schema_pattern=schema_name,
+                table_pattern=table_name,
                 table_types=table_types if table_types else None,
             )
             return self._make_result_set(stream, cursor, self._synthetic_command_id())
@@ -994,10 +961,10 @@ def get_columns(
             # Thrift backend's `getColumns(null, …)` behaviour from
             # the user's perspective.
             stream = self._kernel_session.metadata().list_columns(
-                catalog=_catalog_or_none(catalog_name),
-                schema_pattern=_none_if_blank(schema_name),
-                table_pattern=_none_if_blank(table_name),
-                column_pattern=_none_if_blank(column_name),
+                catalog=catalog_name,
+                schema_pattern=schema_name,
+                table_pattern=table_name,
+                column_pattern=column_name,
             )
             return self._make_result_set(stream, cursor, self._synthetic_command_id())
         except Exception as exc:

src/databricks/sql/client.py

@@ -1577,7 +1577,9 @@ def schemas(
         """
         Get schemas corresponding to the catalog_name and schema_name.
 
-        Names can contain % wildcards.
+        Filters are forwarded unchanged; only ``None`` leaves one unset.
+        ``catalog_name`` is exact. ``schema_name`` is a pattern, can contain
+        % wildcards, and an empty pattern matches nothing.
         :returns self
         """
         self._check_not_closed()
@@ -1603,7 +1605,9 @@ def tables(
         """
         Get tables corresponding to the catalog_name, schema_name and table_name.
 
-        Names can contain % wildcards.
+        Filters are forwarded unchanged; only ``None`` leaves one unset.
+        Names are patterns, can contain % wildcards, and empty patterns match
+        nothing.
         :returns self
         """
         self._check_not_closed()
@@ -1632,7 +1636,9 @@ def columns(
         """
         Get columns corresponding to the catalog_name, schema_name, table_name and column_name.
 
-        Names can contain % wildcards.
+        Filters are forwarded unchanged; only ``None`` leaves one unset.
+        ``catalog_name`` is exact. Other names are patterns, can contain
+        % wildcards, and empty patterns match nothing.
 
         ``catalog_name=None`` is accepted on all backends and matches
         columns across every catalog (the kernel issues ``SHOW COLUMNS``

tests/e2e/test_kernel_backend.py

@@ -356,17 +356,31 @@ def test_metadata_columns(conn):
         assert len(rows) > 0
 
 
-# ── Metadata filter normalization (batch 3) ───────────────────────
+# ── Metadata filter semantics ─────────────────────────────────────
 
 
-def test_schemas_with_empty_string_filter_matches_all(conn):
-    """An empty-string schema pattern normalizes to match-all rather
-    than raising ``ProgrammingError`` (kernel rejects ``""``) — locks
-    ``_none_if_blank`` on the pattern args."""
+def test_schemas_with_empty_string_filter_matches_nothing(conn):
+    """An empty string is a real pattern, distinct from absent ``None``."""
     with conn.cursor() as cur:
         cur.schemas(catalog_name="main", schema_name="")
-        rows = cur.fetchall()
-        assert len(rows) > 0
+        assert cur.fetchall() == []
+
+
+@pytest.mark.parametrize(
+    "empty_filter", ["schema_name", "table_name", "column_name"]
+)
+def test_columns_with_empty_string_filter_matches_nothing(conn, empty_filter):
+    filters = {
+        "catalog_name": "system",
+        "schema_name": "information_schema",
+        "table_name": "tables",
+        "column_name": "table_catalog",
+    }
+    filters[empty_filter] = ""
+
+    with conn.cursor() as cur:
+        cur.columns(**filters)
+        assert cur.fetchall() == []
 
 
 def test_tables_table_types_filter_is_case_insensitive(conn):