Databricks Data Engineer Associate sheet
Last-minute reference for the Databricks Data Engineer Associate exam.
Exam facts
| Questions | 45 scored |
| Duration | 90 minutes |
| Fee | 200 USD |
| Validity | 2 years |
| Passing score | Not published by Databricks |
Section weights
| Section | Weight |
|---|---|
| Data Transformation and Modeling | 22% |
| Data Ingestion and Loading | 21% |
| Working with Lakeflow Jobs | 16% |
| Governance and Security | 15% |
| Implementing CI/CD | 10% |
| Troubleshooting, Monitoring, Optimization | 10% |
| Databricks Intelligence Platform | 6% |
Ingestion plus transformation is 43%.
Pipeline patterns
| Requirement | Pattern |
|---|---|
| Process each new file once, no full listing | Incremental streaming ingestion with checkpointing |
| Re-run after partial failure without duplicates | Idempotent MERGE on a unique key |
| New optional column must not break the write | Schema evolution |
| Apply latest version of each record from a change feed | MERGE INTO on the key |
| Keep attribute history for point-in-time joins | SCD type 2 |
| Null business keys must be investigable | Quality rule routing to a quarantine table |
| Task 3 runs only if 1 and 2 succeed | One job with task dependencies |
| Transient cloud storage errors | Bounded retries with alerting on final failure |
| Start when upstream finishes | Trigger on upstream completion |
Layers
Bronze raw as received · Silver cleansed · Gold business-level aggregates, cheaper and consistently defined for dashboards.
Optimisation
- Filter before the join, then aggregate
- Job suddenly slow with no code change → input volume, small files, skew
- A few tasks dominate a stage → data skew
Governance
- Hide one column → column mask
- Which tables a job reads and writes → lineage in the governance catalog
- Leavers lose access automatically → group-based access
- Promote code with review and traceability → version control plus automated deployment
- Same code, two environments → parameterised catalog, schema and paths
The published-score trap
Databricks publishes no passing score. Treat any quoted percentage as a guess.
Night-before checklist
- Seven sections; transformation and ingestion are 43% together
- Idempotent merge as the correctness answer
- Bronze/silver/gold, cold
- Check ID and proctoring rules — see exam day
Take the 20-question practice test.