This is an automated email from the ASF dual-hosted git repository.
rusackas pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/superset.git
The following commit(s) were added to refs/heads/master by this push:
new 4e09889607f test(datasets): regression coverage for #16141 (export
with same table name, different schemas) (#40123)
4e09889607f is described below
commit 4e09889607f3ca4c697fb683d10b9926200daeba
Author: Evan Rusackas <[email protected]>
AuthorDate: Thu May 14 11:08:23 2026 -0700
test(datasets): regression coverage for #16141 (export with same table
name, different schemas) (#40123)
Co-authored-by: Superset Dev <[email protected]>
Co-authored-by: Claude Opus 4.7 <[email protected]>
---
tests/unit_tests/datasets/commands/export_test.py | 52 +++++++++++++++++++++++
1 file changed, 52 insertions(+)
diff --git a/tests/unit_tests/datasets/commands/export_test.py
b/tests/unit_tests/datasets/commands/export_test.py
index a449b764084..d7cbc9dd37c 100644
--- a/tests/unit_tests/datasets/commands/export_test.py
+++ b/tests/unit_tests/datasets/commands/export_test.py
@@ -18,6 +18,7 @@
from uuid import UUID
+import yaml
from sqlalchemy.orm.session import Session
from superset import db
@@ -304,3 +305,54 @@ version: 1.0.0
""",
),
]
+
+
+def test_export_two_datasets_same_table_name_different_schema(
+ session: Session,
+) -> None:
+ """
+ Regression coverage for GitHub issue #16141.
+
+ Exporting two datasets that share a `table_name` but live in
+ different schemas (e.g. prod.users + dev.users) must produce two
+ distinct entries in the export. Historically the pair could collide
+ onto a single filename — the export filename is now disambiguated by
+ dataset id, so this test pins that behavior so it can't silently
+ regress.
+ """
+ from superset.commands.dataset.export import ExportDatasetsCommand
+ from superset.connectors.sqla.models import SqlaTable
+ from superset.models.core import Database
+
+ engine = db.session.get_bind()
+ SqlaTable.metadata.create_all(engine) # pylint: disable=no-member
+
+ database = Database(database_name="my_database",
sqlalchemy_uri="sqlite://")
+ db.session.add(database)
+ db.session.flush()
+
+ prod = SqlaTable(table_name="users", schema="prod", database=database)
+ dev = SqlaTable(table_name="users", schema="dev", database=database)
+ db.session.add_all([prod, dev])
+ db.session.flush()
+
+ paths: list[str] = []
+ contents: list[str] = []
+ for ds in (prod, dev):
+ for path, content_fn in ExportDatasetsCommand._export( # pylint:
disable=protected-access
+ ds, export_related=False
+ ):
+ paths.append(path)
+ contents.append(content_fn())
+
+ # Both datasets must produce distinct export paths — no collision.
+ assert len(paths) == len(set(paths)), (
+ f"Export filenames collided for same-table-name datasets: {paths}"
+ )
+
+ # And both YAML payloads must reflect their own schema, not be
+ # silently merged or overwritten.
+ schemas_in_yaml = {yaml.safe_load(c)["schema"] for c in contents}
+ assert schemas_in_yaml == {"prod", "dev"}, (
+ f"Expected both prod and dev schemas in export, got {schemas_in_yaml}"
+ )
