You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The dbt catalog functionality doesn't support missing database fields.
In a number of places, dbt expects database to be a str, but we should allow it to be None for some databases that don't have support for databases.
Catalogs should be allowed to set the database to None for data warehouses that don't have them.
list_relations should be allowed to take None for its database
on most data warehouses, None should either be an error or mean target.database.
on Spark, anything but None or the schema would be an error.
get_relation should get the same treatment
That's probably it?
Should this be extended to schemas?
Describe alternatives you've considered
We can set database='' everywhere. Kind of ugly! I would like to be able to differentiate "this component does not exist/has no meaning" vs "this is the empty string"...
Edit: This actually doesn't work without writing a bunch of new agate type tester things because by default, '' is coerced into None (which makes sense for seeds!).
Additional context
This is mostly relevant for Spark.
Who will this benefit?
Spark users, developers.
The text was updated successfully, but these errors were encountered:
Describe the feature
The dbt catalog functionality doesn't support missing database fields.
In a number of places, dbt expects database to be a
str
, but we should allow it to beNone
for some databases that don't have support for databases.None
for its databaseNone
should either be an error or mean target.database.None
or the schema would be an error.Should this be extended to schemas?
Describe alternatives you've considered
We can set
database=''
everywhere. Kind of ugly! I would like to be able to differentiate "this component does not exist/has no meaning" vs "this is the empty string"...Edit: This actually doesn't work without writing a bunch of new agate type tester things because by default,
''
is coerced intoNone
(which makes sense for seeds!).Additional context
This is mostly relevant for Spark.
Who will this benefit?
Spark users, developers.
The text was updated successfully, but these errors were encountered: