I'm trying create a Meltano task with dbt like so:

- dbt-redshift:run -s customerio

When I run this with the

-s

selector I get the following error:

Error: Block -s not found

Something to do with the way Meltano is parsing the YML I think. Adding quotes doesn't help. How can I select the models to run. Doing

dbt-redshift:run

alone works.

Ran into the same thing recently. You would need to use invoke, as jobs by default use

meltano run

. Invoke allows you to pass in plugin specific arguments. Here is an example of a workaround that I have implemented.

utilities
- name: dbt-wrapper
    namespace: dbt-wrappers
    commands:
      run:
        args: invoke dbt-postgres:run --select models/data
        executable: meltano

And then in your task:

- dbt-wrapper:run

OK it works with the wrapper approach, but feels kinda clunky. Can I not define the command under the existing transformer entry like so:

commands:
      run_customerio:
        args: invoke dbt-redshift:run -s customerio
        executable: meltano

This currently gives me an error saying:

Cannot start plugin: Executable 'dbt' could not be found. Transformer 'dbt-redshift' may not have been installed yet using `meltano install transformer dbt-redshift`,

This works for me for running `dbt build --empty`:

utilities:
  - name: dbt-duckdb
    variant: jwills
    pip_url: dbt-core~=1.10.0b1 dbt-duckdb~=1.9.1 duckdb==1.2.2 meltano-dbt-ext~=0.3.0
    config:
      path: $DUCKDB_PATH
    commands:
      dry-run:
        args: build --empty
        executable: dbt

as

meltano run dbt-duckdb:dry-run

We might just need better examples in the Meltano docs and/or in the Hub.