test: make MV schedule and seed_onto_view functional tests rerun-safe (#1581)

## Description

Two functional tests flake in the nightly Integration suites when a
transient warehouse/auth blip kills an attempt mid-test.
`pytest-rerunfailures` retries in the same process **without** tearing
down the class-scoped `project`, so the leftover relation and the
in-place-rewritten model file leak into the retry, which then fails on a
spurious `AssertionError`:

- `TestMaterializedViewScheduleLifecycle::test_full_lifecycle` — `assert
'cron' == 'manual'` (retry re-enters with the MV already switched to
cron).
- `TestSeedOntoView::test_seed_onto_view_is_rejected` — `dbt exit state
did not match expected` (leftover `seed_over_view` relation).

These are **not** code regressions — the same tests pass on a clean run
— but they turn transient infra flakiness into red CI that reads like a
regression during nightly triage.

## Fix

Adopt the existing `RerunSafeMixin`
(`tests/functional/adapter/fixtures.py`) — the codebase-wide convention
already used by ~30 classes. Its autouse, function-scoped
`reset_project_state` fires on every attempt (the class-scoped `project`
survives reruns), restoring the `models` files to their initial content
and dropping the named `relations_to_reset` before the retry, so each
attempt starts from a clean slate. On the first attempt it is a no-op
(files already at initial content, relations don't exist yet), so
**happy-path behavior is unchanged**.

- `test_schedule_modes.py`: all four MV schedule classes inherit
`RerunSafeMixin` with a `relations_to_reset` naming their single MV
(`mv_on_update_bare`, `mv_drop_readd`, `mv_lifecycle`,
`mv_every_inputs`). The observed failure was `mv_lifecycle`; the
siblings share the identical latent leak (in-place `write_file` + fixed
names), so they are hardened too.
- `test_seeds.py`: `TestSeedOntoView` inherits `RerunSafeMixin` with
`relations_to_reset = ("seed_over_view",)`; `TestSeedOntoViewV2`
inherits it transitively.

## Verification

On the SQL-warehouse profile (`databricks_uc_sql_endpoint`):

- **No happy-path regression**: `test_schedule_modes.py` = 4 passed;
`TestSeedOntoView` + `TestSeedOntoViewV2` = 2 passed.
- **Rerun-safety proof** (throwaway test, not committed): injected a
one-shot failure right after the cron transition and ran `--reruns 1`. A
bare no-mixin copy **failed on rerun** with the exact nightly signature
`assert 'cron' == 'manual'`; the `RerunSafeMixin` copy **passed on
rerun**. Final: `1 failed, 1 passed, 2 rerun` — confirming both that the
leak is real and that the mixin fixes it.
- `pre-commit` (ruff, ruff-format, mypy) passes.

Test-only; no adapter/runtime code changes.

## Checklist

- [x] I have run this code in development, and it appears to resolve the
stated issue.
- [x] This PR includes tests, or tests are not required/relevant for
this PR.
- [x] I have updated the `CHANGELOG.md` and added information about my
change.

Author: sd-db

CHANGELOG.md

@@ -30,7 +30,7 @@
 - Raise the `dbt-adapters` upper bound to `<1.25.0` ([#1507](https://github.com/databricks/dbt-databricks/pull/1507))
 - Raise the `databricks-sdk` upper bound to `<0.118.0` to pick up 0.117.0, which fixes `WorkspaceClient` construction failing with `CONTEXT_UNAVAILABLE_FOR_REMOTE_CLIENT` on Spark Connect clusters ([#1517](https://github.com/databricks/dbt-databricks/pull/1517) closes [#1252](https://github.com/databricks/dbt-databricks/issues/1252))
 - Raise the `dbt-core` upper bound to `<1.11.13` to include dbt-core 1.11.12 ([#1578](https://github.com/databricks/dbt-databricks/pull/1578))
-- Make the remaining incremental functional tests (tags, column tags, tblproperties, liquid clustering, column masks, persist_docs, replace table) rerun-safe so a `pytest --reruns` retry no longer inherits mutated state (rewritten `schema.yml`/model files, half-built relations) from the failed attempt (test-only, no runtime impact).
+- Make stateful functional tests rerun-safe so a `pytest --reruns` retry no longer inherits mutated state (rewritten `schema.yml`/model files, half-built relations) from the failed attempt: the remaining incremental tests (tags, column tags, tblproperties, liquid clustering, column masks, persist_docs, replace table), the materialized-view schedule tests, and the `seed_onto_view` test (test-only, no runtime impact). ([#1512](https://github.com/databricks/dbt-databricks/pull/1512) [#1581](https://github.com/databricks/dbt-databricks/pull/1581))
 - Add unit tests for the previously-uncovered live event/logging classes (`QueryError`, `ConnectionCreate`, `ConnectionCreateError`, and `DbtCoreHandler` log-level routing), bringing the live code in `events/` and `logging.py` to full coverage (test-only, no runtime impact). ([#1548](https://github.com/databricks/dbt-databricks/pull/1548))
 
 ## dbt-databricks 1.12.1 (June 10, 2026)

tests/functional/adapter/materialized_view_tests/test_schedule_modes.py

@@ -9,6 +9,7 @@
     MaterializedViewConfig,
 )
 from dbt.adapters.databricks.relation_configs.refresh import RefreshConfig
+from tests.functional.adapter.fixtures import RerunSafeMixin
 from tests.functional.adapter.materialized_view_tests import fixtures
 
 
@@ -27,7 +28,7 @@ def _get_refresh_config(project, identifier):
 
 @pytest.mark.dlt
 @pytest.mark.skip_profile("databricks_cluster", "databricks_uc_cluster")
-class TestMaterializedViewScheduleModes:
+class TestMaterializedViewScheduleModes(RerunSafeMixin):
     @pytest.fixture(scope="class", autouse=True)
     def seeds(self):
         yield {"my_seed.csv": MY_SEED}
@@ -36,6 +37,10 @@ def seeds(self):
     def models(self):
         yield {"mv_on_update_bare.sql": fixtures.materialized_view_on_update_bare}
 
+    @pytest.fixture(scope="class")
+    def relations_to_reset(self):
+        return ("mv_on_update_bare",)
+
     def test_on_update_bare_mode_roundtrip(self, project):
         util.run_dbt(["seed"])
         util.run_dbt(["run", "--models", "mv_on_update_bare"])
@@ -46,7 +51,7 @@ def test_on_update_bare_mode_roundtrip(self, project):
 
 @pytest.mark.dlt
 @pytest.mark.skip_profile("databricks_cluster", "databricks_uc_cluster")
-class TestMaterializedViewDropAndReadd:
+class TestMaterializedViewDropAndReadd(RerunSafeMixin):
     """Drop schedule (config removed) and re-add."""
 
     @pytest.fixture(scope="class", autouse=True)
@@ -57,6 +62,10 @@ def seeds(self):
     def models(self):
         yield {"mv_drop_readd.sql": fixtures.materialized_view_cron_no_tz}
 
+    @pytest.fixture(scope="class")
+    def relations_to_reset(self):
+        return ("mv_drop_readd",)
+
     @pytest.fixture(scope="class")
     def project_config_update(self):
         return {"models": {"on_configuration_change": "apply"}}
@@ -82,7 +91,7 @@ def test_drop_then_readd(self, project):
 
 @pytest.mark.dlt
 @pytest.mark.skip_profile("databricks_cluster", "databricks_uc_cluster")
-class TestMaterializedViewScheduleLifecycle:
+class TestMaterializedViewScheduleLifecycle(RerunSafeMixin):
     """Walks one MV through the realistic schedule lifecycle:
     MANUAL → CRON → ON_UPDATE rate-limited → EVERY → (non-refresh change) → MANUAL.
     Each transition asserts the post-state schedule via DESCRIBE EXTENDED."""
@@ -95,6 +104,10 @@ def seeds(self):
     def models(self):
         yield {"mv_lifecycle.sql": fixtures.materialized_view_no_schedule}
 
+    @pytest.fixture(scope="class")
+    def relations_to_reset(self):
+        return ("mv_lifecycle",)
+
     @pytest.fixture(scope="class")
     def project_config_update(self):
         return {"models": {"on_configuration_change": "apply"}}
@@ -144,7 +157,7 @@ def test_full_lifecycle(self, project):
 
 @pytest.mark.dlt
 @pytest.mark.skip_profile("databricks_cluster", "databricks_uc_cluster")
-class TestMaterializedViewEveryAcceptedInputs:
+class TestMaterializedViewEveryAcceptedInputs(RerunSafeMixin):
     """Walk one MV through each EVERY unit the regex accepts (HOUR / DAY / WEEK) and
     confirm dbt round-trips each against a real warehouse. Uses
     `on_configuration_change: apply` so each iteration is an ALTER, not a fresh CREATE."""
@@ -157,6 +170,10 @@ def seeds(self):
     def models(self):
         yield {"mv_every_inputs.sql": fixtures.materialized_view_with_every("2 HOURS")}
 
+    @pytest.fixture(scope="class")
+    def relations_to_reset(self):
+        return ("mv_every_inputs",)
+
     @pytest.fixture(scope="class")
     def project_config_update(self):
         return {"models": {"on_configuration_change": "apply"}}

tests/functional/adapter/simple_seed/test_seeds.py

@@ -17,7 +17,7 @@
     SeedTestBase,
 )
 
-from tests.functional.adapter.fixtures import MaterializationV2Mixin
+from tests.functional.adapter.fixtures import MaterializationV2Mixin, RerunSafeMixin
 from tests.functional.adapter.simple_seed import fixtures
 
 
@@ -185,7 +185,11 @@ class TestSeedColumnTypesV2(TestSeedColumnTypes, MaterializationV2Mixin):
     pass
 
 
-class TestSeedOntoView:
+class TestSeedOntoView(RerunSafeMixin):
+    @pytest.fixture(scope="class")
+    def relations_to_reset(self):
+        return ("seed_over_view",)
+
     @pytest.fixture(scope="class")
     def seeds(self):
         return {"seed_over_view.csv": fixtures.seeds__over_view_csv}