I’m working on a dagster POC and having some trouble adapting to the asset-based model. I’m trying to just run a simple fivetran sync and then run a SQL query to create a table based on this data. Is there a way to define an asset as a sql query to run (in this case it’s a merge statement) - I got it to run succesfully as an op, but am unsure how I can pass that downstream to something

Hey Sam, there are couple of ways to do this: 1. If you want the SQL transformation to happen in your Snowflake warehouse, you can have dbt run downstream of your fivetran sync. We have a guide for setting setting up dbt here: https://docs.dagster.io/integrations/dbt 2. If you want to do the transformation without dbt, you can configure the

dagster-snowflake

resource and define an asset that runs the one-off sql query in your data warehouse We probably recommend doing (1), if you’re planning on implementing more transformations of your Fivetran synced data in the future.

due to a few process oriented things, we actually can’t use dbt to do this, so I need to go with option 2.

Re: requiring assets upstream: https://docs.dagster.io/concepts/assets/software-defined-assets#non-argument-dependencies In this case, you are not passing data between your assets, since the transformation happens entirely in your Snowflake warehouse. So you should be using

non_argument_deps

in your downstream asset.