unit test error: jsonschema.exceptions.ValidationError: 'database' is a required property

[dataders]

Background

somehow, dbt-labs/dbt-core#8437 has caused a dbt-databricks unit test to fail with the below strange message. Here's a gist of the failure message I saw locally (TODO add link to GHA tox log)

jsonschema.exceptions.ValidationError: 'database' is a required property

I’m confused bc, AFAICT, database is defined as optional within the DatabricksCredentials dataclass.

dbt-databricks/dbt/adapters/databricks/connections.py

Lines 102 to 103 in 3609a6e

	class DatabricksCredentials(Credentials):
	database: Optional[str] # type: ignore[assignment]

It seems as if I was able to fix this by adding "database": os.getenv("DBT_DATABRICKS_DB_NAME"), to the profile dict definition

dbt-databricks/tests/profiles.py

Lines 23 to 26 in 3609a6e

	"host": os.getenv("DBT_DATABRICKS_HOST_NAME"),
	"http_path": http_path,
	"token": os.getenv("DBT_DATABRICKS_TOKEN"),
	"client_id": os.getenv("DBT_DATABRICKS_CLIENT_ID"),

Adding to my confusion is that the failing unit test, TestDatabricksAdapter.test_databricks_sql_connector_connection, looks to based upon dbt-spark’s TestSparkAdapter.test_http_connection() unit test. Both tests are of a format unfamiliar to me and structured differently than dbt-snowflake's TestSnowflakeAdapter, for example.

Questions

1. where in the unit test is profile validation happening?
2. why did this not happen before?
3. are the unit tests of dbt-databricks (therefore dbt-spark) out-of-date, or not best practice?

[dataders]

nevermind!!!