Databricks Asset Bundles (DABs) Variables

Databricks Asset Bundles allow you to express your data projects (jobs, pipelines, infrastructure) as code. A key challenge in Infrastructure as Code (IaC) is managing configuration drift between environments (Dev, Staging, Prod). DABs solves this with Variables and Targets.

Targets: Represents an environment (e.g., dev, prod). You can define specific overrides for each target.
Variables: defined in the variables section.
You can set a default value.
You can override the value inside a specific target.

Usage example in databricks.yml:

variables:
  schema_name:
    description: 'The schema to write to'
    default: 'default_schema'
 
targets:
  prod:
    variables:
      schema_name: 'prod_schema'
 
resources:
  jobs:
    my_job:
      tasks:
        - task_key: t1
          spark_python_task:
            parameters:
              - '${var.schema_name}'

When you run databricks bundle deploy -t prod, the value passed to the job will be "prod_schema".

Topic: Jobs