This is an automated email from the ASF dual-hosted git repository.
fokko pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/iceberg-python.git
The following commit(s) were added to refs/heads/main by this push:
new 33bfbc88 Add `namespace_exists` method (#2972)
33bfbc88 is described below
commit 33bfbc8807279c77563351ec1c4d0b044871440b
Author: Alex Stephen <[email protected]>
AuthorDate: Wed Jan 28 13:17:05 2026 -0800
Add `namespace_exists` method (#2972)
<!--
Thanks for opening a pull request!
-->
<!-- In the case this PR will resolve an issue, please replace
${GITHUB_ISSUE_ID} below with the actual Github issue id. -->
Closes #2969
# Rationale for this change
We should make `namespace_exists` available on every catalog.
## Are these changes tested?
Includes unit tests. Many catalogs have a private method doing this
already. Look at all of the type exceptions being deleted!
## Are there any user-facing changes?
- Adds namespace_exists method.
<!-- In the case of user-facing changes, please add the changelog label.
-->
---
pyiceberg/catalog/__init__.py | 26 ++++++++++++++++++++++++++
pyiceberg/catalog/noop.py | 3 +++
pyiceberg/catalog/sql.py | 20 ++++++++++----------
tests/catalog/test_catalog_behaviors.py | 24 ++++++++++++------------
tests/integration/test_catalog.py | 3 +++
5 files changed, 54 insertions(+), 22 deletions(-)
diff --git a/pyiceberg/catalog/__init__.py b/pyiceberg/catalog/__init__.py
index 67ca805c..31552155 100644
--- a/pyiceberg/catalog/__init__.py
+++ b/pyiceberg/catalog/__init__.py
@@ -470,6 +470,17 @@ class Catalog(ABC):
bool: True if the view exists, False otherwise.
"""
+ @abstractmethod
+ def namespace_exists(self, namespace: str | Identifier) -> bool:
+ """Check if a namespace exists.
+
+ Args:
+ namespace (str | Identifier): Namespace identifier.
+
+ Returns:
+ bool: True if the namespace exists, False otherwise.
+ """
+
@abstractmethod
def register_table(self, identifier: str | Identifier, metadata_location:
str) -> Table:
"""Register a new table using existing metadata.
@@ -845,6 +856,21 @@ class MetastoreCatalog(Catalog, ABC):
except NoSuchTableError:
return False
+ def namespace_exists(self, namespace: str | Identifier) -> bool:
+ """Check if a namespace exists.
+
+ Args:
+ namespace (str | Identifier): Namespace identifier.
+
+ Returns:
+ bool: True if the namespace exists, False otherwise.
+ """
+ try:
+ self.load_namespace_properties(namespace)
+ return True
+ except NoSuchNamespaceError:
+ return False
+
def purge_table(self, identifier: str | Identifier) -> None:
table = self.load_table(identifier)
self.drop_table(identifier)
diff --git a/pyiceberg/catalog/noop.py b/pyiceberg/catalog/noop.py
index ac2423c1..0dc6fdb7 100644
--- a/pyiceberg/catalog/noop.py
+++ b/pyiceberg/catalog/noop.py
@@ -122,5 +122,8 @@ class NoopCatalog(Catalog):
def view_exists(self, identifier: str | Identifier) -> bool:
raise NotImplementedError
+ def namespace_exists(self, namespace: str | Identifier) -> bool:
+ raise NotImplementedError
+
def drop_view(self, identifier: str | Identifier) -> None:
raise NotImplementedError
diff --git a/pyiceberg/catalog/sql.py b/pyiceberg/catalog/sql.py
index a2704c3d..dc24abcf 100644
--- a/pyiceberg/catalog/sql.py
+++ b/pyiceberg/catalog/sql.py
@@ -204,7 +204,7 @@ class SqlCatalog(MetastoreCatalog):
namespace_identifier = Catalog.namespace_from(identifier)
table_name = Catalog.table_name_from(identifier)
- if not self._namespace_exists(namespace_identifier):
+ if not self.namespace_exists(namespace_identifier):
raise NoSuchNamespaceError(f"Namespace does not exist:
{namespace_identifier}")
namespace = Catalog.namespace_to_string(namespace_identifier)
@@ -251,7 +251,7 @@ class SqlCatalog(MetastoreCatalog):
namespace_tuple = Catalog.namespace_from(identifier)
namespace = Catalog.namespace_to_string(namespace_tuple)
table_name = Catalog.table_name_from(identifier)
- if not self._namespace_exists(namespace):
+ if not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace does not exist:
{namespace}")
with Session(self.engine) as session:
@@ -361,7 +361,7 @@ class SqlCatalog(MetastoreCatalog):
to_namespace_tuple = Catalog.namespace_from(to_identifier)
to_namespace = Catalog.namespace_to_string(to_namespace_tuple)
to_table_name = Catalog.table_name_from(to_identifier)
- if not self._namespace_exists(to_namespace):
+ if not self.namespace_exists(to_namespace):
raise NoSuchNamespaceError(f"Namespace does not exist:
{to_namespace}")
with Session(self.engine) as session:
try:
@@ -495,7 +495,7 @@ class SqlCatalog(MetastoreCatalog):
metadata=updated_staged_table.metadata,
metadata_location=updated_staged_table.metadata_location
)
- def _namespace_exists(self, identifier: str | Identifier) -> bool:
+ def namespace_exists(self, identifier: str | Identifier) -> bool:
namespace_tuple = Catalog.identifier_to_tuple(identifier)
namespace = Catalog.namespace_to_string(namespace_tuple,
NoSuchNamespaceError)
namespace_starts_with = namespace.replace("!", "!!").replace("_",
"!_").replace("%", "!%") + ".%"
@@ -537,7 +537,7 @@ class SqlCatalog(MetastoreCatalog):
Raises:
NamespaceAlreadyExistsError: If a namespace with the given name
already exists.
"""
- if self._namespace_exists(namespace):
+ if self.namespace_exists(namespace):
raise NamespaceAlreadyExistsError(f"Namespace {namespace} already
exists")
if not properties:
@@ -565,7 +565,7 @@ class SqlCatalog(MetastoreCatalog):
NoSuchNamespaceError: If a namespace with the given name does not
exist.
NamespaceNotEmptyError: If the namespace is not empty.
"""
- if not self._namespace_exists(namespace):
+ if not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace does not exist:
{namespace}")
namespace_str = Catalog.namespace_to_string(namespace)
@@ -593,7 +593,7 @@ class SqlCatalog(MetastoreCatalog):
Raises:
NoSuchNamespaceError: If a namespace with the given name does not
exist.
"""
- if namespace and not self._namespace_exists(namespace):
+ if namespace and not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace does not exist:
{namespace}")
namespace = Catalog.namespace_to_string(namespace)
@@ -614,7 +614,7 @@ class SqlCatalog(MetastoreCatalog):
Raises:
NoSuchNamespaceError: If a namespace with the given name does not
exist.
"""
- if namespace and not self._namespace_exists(namespace):
+ if namespace and not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace does not exist:
{namespace}")
table_stmt =
select(IcebergTables.table_namespace).where(IcebergTables.catalog_name ==
self.name)
@@ -656,7 +656,7 @@ class SqlCatalog(MetastoreCatalog):
NoSuchNamespaceError: If a namespace with the given name does not
exist.
"""
namespace_str = Catalog.namespace_to_string(namespace)
- if not self._namespace_exists(namespace):
+ if not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace {namespace_str} does not
exists")
stmt = select(IcebergNamespaceProperties).where(
@@ -681,7 +681,7 @@ class SqlCatalog(MetastoreCatalog):
ValueError: If removals and updates have overlapping keys.
"""
namespace_str = Catalog.namespace_to_string(namespace)
- if not self._namespace_exists(namespace):
+ if not self.namespace_exists(namespace):
raise NoSuchNamespaceError(f"Namespace {namespace_str} does not
exists")
current_properties =
self.load_namespace_properties(namespace=namespace)
diff --git a/tests/catalog/test_catalog_behaviors.py
b/tests/catalog/test_catalog_behaviors.py
index 29689358..21508a8f 100644
--- a/tests/catalog/test_catalog_behaviors.py
+++ b/tests/catalog/test_catalog_behaviors.py
@@ -946,7 +946,7 @@ def test_add_column_with_statement(catalog: Catalog,
table_schema_simple: Schema
def test_create_namespace(catalog: Catalog, test_namespace: Identifier,
test_table_properties: dict[str, str]) -> None:
catalog.create_namespace(test_namespace, test_table_properties)
- assert catalog._namespace_exists(test_namespace) # type:
ignore[attr-defined]
+ assert catalog.namespace_exists(test_namespace)
assert (Catalog.identifier_to_tuple(test_namespace)[:1]) in
catalog.list_namespaces()
assert test_table_properties ==
catalog.load_namespace_properties(test_namespace)
@@ -1015,16 +1015,16 @@ def
test_get_namespace_metadata_raises_error_when_namespace_does_not_exist(catal
def test_namespace_exists(catalog: Catalog) -> None:
for ns in [("db1",), ("db1", "ns1"), ("db2", "ns1"), ("db3", "ns1",
"ns2")]:
catalog.create_namespace(ns)
- assert catalog._namespace_exists(ns) # type: ignore[attr-defined]
+ assert catalog.namespace_exists(ns)
# `db2` exists because `db2.ns1` exists
- assert catalog._namespace_exists("db2") # type: ignore[attr-defined]
+ assert catalog.namespace_exists("db2")
# `db3.ns1` exists because `db3.ns1.ns2` exists
- assert catalog._namespace_exists("db3.ns1") # type: ignore[attr-defined]
+ assert catalog.namespace_exists("db3.ns1")
# make sure '_' is escaped in the query
- assert not catalog._namespace_exists("db_") # type: ignore[attr-defined]
+ assert not catalog.namespace_exists("db_")
# make sure '%' is escaped in the query
- assert not catalog._namespace_exists("db%") # type: ignore[attr-defined]
+ assert not catalog.namespace_exists("db%")
# Namespace properties
@@ -1064,7 +1064,7 @@ def test_load_empty_namespace_properties(catalog:
Catalog, test_namespace: Ident
def test_list_namespaces(catalog: Catalog) -> None:
namespace_list = ["db", "db.ns1", "db.ns1.ns2", "db.ns2", "db2",
"db2.ns1", "db%"]
for namespace in namespace_list:
- if not catalog._namespace_exists(namespace): # type:
ignore[attr-defined]
+ if not catalog.namespace_exists(namespace):
catalog.create_namespace(namespace)
ns_list = catalog.list_namespaces()
@@ -1084,7 +1084,7 @@ def test_list_namespaces(catalog: Catalog) -> None:
def test_list_namespaces_fuzzy_match(catalog: Catalog) -> None:
namespace_list = ["db.ns1", "db.ns1.ns2", "db.ns2", "db.ns1X.ns3",
"db_.ns1.ns2", "db2.ns1.ns2"]
for namespace in namespace_list:
- if not catalog._namespace_exists(namespace): # type:
ignore[attr-defined]
+ if not catalog.namespace_exists(namespace):
catalog.create_namespace(namespace)
assert catalog.list_namespaces("db.ns1") == [("db", "ns1", "ns2")]
@@ -1140,7 +1140,7 @@ def test_update_namespace_metadata(catalog: Catalog,
test_namespace: Identifier,
catalog.create_namespace(test_namespace, test_table_properties)
new_metadata = {"key3": "value3", "key4": "value4"}
summary = catalog.update_namespace_properties(test_namespace,
updates=new_metadata)
- assert catalog._namespace_exists(test_namespace) # type:
ignore[attr-defined]
+ assert catalog.namespace_exists(test_namespace)
assert new_metadata.items() <=
catalog.load_namespace_properties(test_namespace).items()
assert summary.removed == []
assert sorted(summary.updated) == ["key3", "key4"]
@@ -1154,7 +1154,7 @@ def test_update_namespace_metadata_removals(
new_metadata = {"key3": "value3", "key4": "value4"}
remove_metadata = {"key1"}
summary = catalog.update_namespace_properties(test_namespace,
remove_metadata, new_metadata)
- assert catalog._namespace_exists(test_namespace) # type:
ignore[attr-defined]
+ assert catalog.namespace_exists(test_namespace)
assert new_metadata.items() <=
catalog.load_namespace_properties(test_namespace).items()
assert
remove_metadata.isdisjoint(catalog.load_namespace_properties(test_namespace).keys())
assert summary.removed == ["key1"]
@@ -1168,13 +1168,13 @@ def test_update_namespace_metadata_removals(
def test_drop_namespace(catalog: Catalog, table_schema_nested: Schema,
test_table_identifier: Identifier) -> None:
namespace = Catalog.namespace_from(test_table_identifier)
catalog.create_namespace(namespace)
- assert catalog._namespace_exists(namespace) # type: ignore[attr-defined]
+ assert catalog.namespace_exists(namespace)
catalog.create_table(test_table_identifier, table_schema_nested)
with pytest.raises(NamespaceNotEmptyError):
catalog.drop_namespace(namespace)
catalog.drop_table(test_table_identifier)
catalog.drop_namespace(namespace)
- assert not catalog._namespace_exists(namespace) # type:
ignore[attr-defined]
+ assert not catalog.namespace_exists(namespace)
def test_drop_namespace_raises_error_when_namespace_does_not_exist(catalog:
Catalog) -> None:
diff --git a/tests/integration/test_catalog.py
b/tests/integration/test_catalog.py
index 72ef16cf..f9d174dd 100644
--- a/tests/integration/test_catalog.py
+++ b/tests/integration/test_catalog.py
@@ -410,7 +410,10 @@ def test_concurrent_create_transaction(test_catalog:
Catalog, test_schema: Schem
@pytest.mark.integration
@pytest.mark.parametrize("test_catalog", CATALOGS)
def test_create_namespace(test_catalog: Catalog, database_name: str) -> None:
+ assert not test_catalog.namespace_exists(database_name)
+
test_catalog.create_namespace(database_name)
+ assert test_catalog.namespace_exists(database_name)
assert (database_name,) in test_catalog.list_namespaces()
