Hi, I'm testing out the new way of using assets. Which so far looks great! My question: How can I add the target and other variables in the define_asset_job? Previous I have used two different configurations for dev and prod jobs in where I set the target that is defined in my profiles.yml

# Dev configuration 
config_dbt_datamart_group_run_dev = {"models": ["DBT_DATAMART_GROUP_RUN"],
"profiles_dir": str(profile_path),
"project-dir": str(Path(curdir) / 'EDW_DQ'),
"vars" : {"prod_source_db": "EDW_DEV", "prod_source_schema":"DATAMART_GROUP"},
"target" : "dev"}  

@job(resource_defs={"dbt": dbt_cli_resource.configured(config_dbt_datamart_group_run_dev)}, 
	name='job_dbt_datamart_group_run_dev', description='This job loads the whole DATAMART_GROUP model  ')
	def job_dbt_datamart_group_run_dev():
    dbt_run_op()

The new version: I load all assets and then in the job. Our DBT project contains several different models. I filter on the specific model in the selection part.

# Assets

dbt_odp_assets = with_resources(
    load_assets_from_dbt_project(
    DBT_PROJECT_DIR,
    DBT_PROFILES_DIR,
    source_key_prefix =["snowflake"]
    ),
    {
        "dbt": dbt_cli_resource.configured(
            {"project_dir": DBT_PROJECT_DIR, "profiles_dir": DBT_PROFILES_DIR}
        )
    },
)

#Jobs

job_datamart_group_dev = define_asset_job(name="run_datamart_group_dev", 
						selection="QUALITY/DATAMART_GROUP_ERROR_EVENT_FACT_AGG*", 
						description="This job loads the whole DATAMART_GROUP model ")

Or is there a better way to do it?

Hi Mikael! We recently released a local development to production guide that describes how you can structure your assets and resources to operate with different configurations for multiple environments. It's similar to what you wrote, but we suggest a resource mapping by environment using the configured API. Something like this:

@repository
def repo():
    resource_defs = {
        "local": {
            "snowflake_io_manager": snowflake_io_manager.configured(
                {
                    ...
                }
            ),
        },
        "production": {
            "snowflake_io_manager": snowflake_io_manager.configured(
                {
                    ...
                }
            ),
        },
    }
    deployment_name = os.getenv("DAGSTER_DEPLOYMENT", "local")

    return [
        *with_resources(
            [items, comments, stories], resource_defs=resource_defs[deployment_name]
        )
    ]

Thank you Claire! Excellent! Sorry that I missed the existing guide. However, This new release of Dagster is great!