dbt run --select spark_test.sql --profile sparkprofile checks presence of "database" even in sources that are not referred in model

[soham-dasgupta]

dbt-core 1.8.9
spark 1.8.0

In addition to using Trino and Athena in dbt, we would like to use Spark and here are the relevant configurations

sources.yml

  - name: spark
    description: glue default database
    schema: default
    tables:
      - name: fact_customer_item_dly
        description: fact_customer_item_dly

profiles.yml

sparkprofile:
  target: dev
  outputs:
    dev:
      type: spark
      method: thrift
      host: localhost
      port: 10001
      schema: dev

spark_test.sql

{{ config(
    materialized='table',
    file_format='parquet'
) }}


select count(*) from {{ source('spark' ,'fact_customer_item_dly') }}

We now want to run the model spark_test.sql and invoke the below command

.venv/bin/dbt run --select spark_test.sql --profile sparkprofile

However we encounter the below error

Runtime Error
  Cannot set database in spark!

The model executes fine if I remove database from sources that are not related to Spark(Trino and Athena). This limits us in using Spark in addition to Trino and Athena, How can we solve this issue?

[soham-dasgupta]

Issue created in dbt adapters repo dbt-labs/dbt-adapters#849

[data-blade]

we worked around (in our case having databricks/spark dual adapter) via

sources:
  - name: source_a
    database: "{{ none if target.type == 'spark' else 'catalog_a' }}"
    schema: "{{ 'catalog_a.schema_a' if target.type == 'spark' else 'schema_a' }}"