Free to read · Three levels
dbt: model contracts and recoverable analytics
A complete dbt learning path connecting SQL model meaning, graph dependencies, tests, incremental history and release evidence.
Start learningIn this course
Develop analytics for a synthetic regional library through eighteen complete lessons. Follow loans, branch references, return corrections, source delivery and daily reports while making every model’s grain, input and publication contract explicit.
Progress from sources, refs, Jinja, materializations and failing-row tests to contracts, snapshots, incremental transitions, unit fixtures, state-aware CI and operational recovery. Every lesson includes developed explanations, worked examples, contextual mistakes, exercises and official primary references, with three stand-alone troubleshooting guides.
What you will learn
- Define source and model contracts with precise grain, references, types and business assertions.
- Design replay-aware incremental models, observed snapshots, governed seeds and focused unit fixtures.
- Review selectors, artifacts, CI isolation, packages, hooks and product-generation boundaries using scoped evidence.
Your learning path
Three levels
Basic: project and model foundations
Define framework boundaries, logical inputs, model grain, template context and tested assertions.
- Framework, runtime and project boundaries
- Sources, staging and dependency graphs
- Model grain, types and business SQL
- Jinja, macros and execution context
- Materializations and configuration precedence
- Data tests and failing records
Intermediate: history and controlled workflows
Build stable interfaces and explicit state transitions with governed fixtures, reference data and selection.
- Contracts, documentation and model interfaces
- Incremental correctness and late changes
- Snapshots and observed history
- Seeds, reference data and source freshness
- Selectors and controlled execution scope
- Unit tests and fixture boundaries
Advanced: evidence, isolation and recoverable releases
Review artifacts, deferred CI, target cost, dependencies, side effects and generation changes.
- Artifacts, lineage and audit evidence
- CI, deferral and environment isolation
- Performance evidence and target query cost
- Packages, dispatch and adapter compatibility
- Orchestration, hooks and partial failures
- Release migration and product boundaries
Who this is for
- Analytics engineers and developers learning dbt with explicit data meaning and execution boundaries.
- Reviewers responsible for model quality, dependency scope, incremental correctness and safe release evidence.
Before you start
- Basic SQL, tabular data, version control and familiarity with YAML-like configuration.
- Optional framework practice requires a separately authorized disposable project and supported adapter/target; this course provisions no warehouse or service.