Open timvw opened 1 year ago
Thanks for raising this idea @timvw 💡 !
Would you be willing to share the code for your location_clause
macro in addition to examples using it? And also an example of how you are proposing to specify the location directly instead?
That way, we can take a look at your proposal vs. the alternative you are using currently.
The existing implementation of location_clause and the difference with my proposal can be found here: https://github.com/dbt-labs/dbt-spark/pull/835/files
Usage is straightforward, simply add a location parameter to your model config:
models:
- name: model_a
config:
format: delta
location: 's3://bucket/sourcesystem/conformed/model-a'
...
Without this feature we would be forced to rename our model to model-a..
Thanks for this example and explanation @timvw !
What do you think @dataders? Do you see this as tying into https://github.com/dbt-labs/dbt-core/discussions/8617 at all?
This issue has been marked as Stale because it has been open for 180 days with no activity. If you would like the issue to remain open, please comment on the issue or else it will be closed in 7 days.
Is this your first time submitting a feature request?
Describe the feature
Currently we have a location_clause macro in place which calculates the location where the data should be stored as
{{ location_root }}/{{ identifier }}
.It would be nice to be able to specify location directly (independent of
location_root
andidentifier
). Eg: we have naming conventions where the tables are nameblah_blah
and locations ares3://bucket/blah-blah
.Describe alternatives you've considered
We could not implement this and push this to user-space
Who will this benefit?
Anyone that wants to specify a location that is not the result of concatenating
location_root
andidentifier
.Are you interested in contributing this feature?
Yes
Anything else?