The Databricks Certified Data Engineer Professional exam currently has 59 scored multiple-choice questions with a 120-minute limit, costs USD 200 plus applicable taxes, and requires no formal prerequisite. It covers ten areas, from developing Python and SQL for data processing (22%) through data modeling (6%). Prepare by reading the official exam guide outline, taking the recommended training, drilling your weak sections, and re-checking the guide about two weeks before your exam date, since Databricks updates it whenever the exam changes.
What the exam covers and which version applies
The exam validates advanced skills in building, optimizing, and maintaining production-grade data engineering solutions on the Databricks platform. The current exam guide is the version live as of July 3, 2026.
Databricks describes this as a professional-level certification. It targets people who design secure, reliable, and cost-effective ETL pipelines, process complex data from diverse sources using Python and SQL, and apply best practices in schema management, observability, governance, and performance optimization.
The platform features named in the current guide are Delta Lake, Unity Catalog, Auto Loader, Lakeflow Spark Declarative Pipelines, Databricks Compute including serverless, Lakeflow Jobs, and the Medallion Architecture. You are also tested on streaming workloads, workflow orchestration, DevOps and CI/CD, and deployment with the Databricks CLI, REST API, and bundles.
Tooling names in this certification have been renamed recently, so anchor your vocabulary to the current guide. The guide itself notes the former names for some of them: Declarative Automation Bundles was previously called Databricks Asset Bundles, AUTO CDC APIs were previously APPLY CHANGES, and Databricks Git folders were previously Repos. Lakeflow Spark Declarative Pipelines is the current name for the declarative pipeline feature covered throughout the outline.
| Testing area | Weight | What the outline emphasizes |
|---|---|---|
| Developing code using Python and SQL | 22% | Project structure for Declarative Automation Bundles, third-party library dependencies, Pandas/Python UDFs, building and testing ETL pipelines, streaming tables vs. materialized views, AUTO CDC APIs, control flow, unit and integration testing |
| Data ingestion & acquisition | 7% | Ingesting Delta, Parquet, ORC, AVRO, JSON, CSV, XML, Text, and Binary from buses and cloud storage; append-only pipelines handling batch and streaming |
| Transformation, cleansing, and quality | 10% | Window functions, joins, aggregations; quarantining bad data with declarative pipelines or Auto Loader in classic jobs |
| Data sharing and federation | 5% | Delta Sharing Databricks-to-Databricks (D2D) and open protocol (D2O); Lakehouse Federation with governance |
| Monitoring and alerting | 10% | System tables, Query Profiler UI, Spark UI, REST API/CLI, pipeline event logs; SQL Alerts and Jobs notifications |
| Cost & performance optimization | 13% | Unity Catalog managed tables, deletion vectors, liquid clustering, data skipping and file pruning, Change Data Feed, query profiles |
| Security and compliance | 10% | ACLs and least privilege, row filters and column masks, hashing, tokenization, suppression, generalization; PII masking and data purging |
| Data governance | 7% | Metadata and descriptions for discoverability; Unity Catalog permission inheritance |
| Debugging and deploying | 10% | Spark UI, cluster logs, system tables, query profiles; job repairs and parameter overrides; Declarative Automation Bundles and Git folders for CI/CD |
| Data modeling | 6% | Scalable Delta Lake models, liquid clustering vs. partitioning and Z-Order, dimensional models |
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; Databricks Certified Data Engineer Professional | Databricks — Official certification page
Format, timing, and cost
59 scored multiple-choice questions, 120 minutes, USD 200 plus applicable taxes, delivered online or at a test center with no test aids.
The exam contains 59 scored multiple-choice questions with a 120-minute time limit. Registration costs USD 200, plus taxes where local law requires them. You can take it online proctored or at a test center, and no test aids are provided, including API documentation.
Databricks states that exams may include unscored items used to gather statistics for future use. These are not identified on the form, do not affect your score, and extra time is already factored in to cover them.
There is no formal prerequisite. Databricks recommends related course attendance and about one year of hands-on experience with the data engineering tasks in the exam guide. Code examples on the exam are primarily in Python and SQL.
The credential is valid for two years. To keep your certified status, you recertify every two years by taking the current version of the exam.
- Questions: 59 scored multiple-choice
- Time limit: 120 minutes
- Fee: USD 200 plus applicable local taxes
- Delivery: online proctored or test center
- Test aids: none, including API documentation
- Prerequisite: none; one year of hands-on experience recommended
- Validity: 2 years, recertify by taking the current exam
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; Databricks Certified Data Engineer Professional | Databricks — Official certification page
Worked example: select one latest event per customer
A window function can rank events within each customer. A complete ordering makes the chosen row predictable.
The exam outline includes window functions under data transformation. Databricks documents that row_number assigns a unique sequence within a partition, beginning at one; its ordering must be unique for deterministic results. QUALIFY filters window-function results.
Practice
Original paper exercise: assume event_id is unique within each customer, has an integer type, and event_time has a timestamp type. The events table contains only the three rows above. Which event IDs remain after this Databricks SQL query? SELECT customer_id, event_id FROM events QUALIFY row_number() OVER (PARTITION BY customer_id ORDER BY event_time DESC, event_id DESC) = 1; Why include event_id in the ordering?
Show answer
C1 retains event18 and C2 retains event9. For C1, both timestamps match, so descending event_id selects18 before17. The query does not specify an order for the final displayed rows.
PARTITION BY restarts the sequence for each customer. Ordering by timestamp alone would leave the C1 tie unresolved; row_number still gives one row number1, but the choice would not be deterministic. The unique event_id breaks that tie. QUALIFY retains the row numbered1 in each partition. This is an original learning example, not an official or recalled exam question.
| customer_id | event_id | event_time |
|---|---|---|
| C1 | 17 | 2026-09-10 10:00:00 |
| C1 | 18 | 2026-09-10 10:00:00 |
| C2 | 9 | 2026-09-10 09:00:00 |
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; row_number ranking window function | Databricks on AWS — Assigns a unique, sequential number; QUALIFY clause | Databricks on AWS — https://docs.databricks.com/aws/en/sql/language-manual/sql-ref-syntax-qry-select-qualify
Use the weights to find gaps in your practice
Compare your experience against every domain, then spend extra practice time where your understanding is weakest.
The published weights show the content balance. They do not tell you the exact questions on your form or how much study time you personally need. Cover smaller domains as well as developing code and performance optimization.
As a study method, keep a checklist with the objective, a task you can complete in a disposable lab, and the result you checked. For example, window-function practice should include ties and partitions, rather than memorizing a query that works for only one set of rows.
Use the retired sample questions linked in the official guide to understand its question format. Check explanations against current documentation when platform behavior matters.
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; Databricks Certified Data Engineer Professional | Databricks — Official certification page; row_number ranking window function | Databricks on AWS — Assigns a unique, sequential number; QUALIFY clause | Databricks on AWS — https://docs.databricks.com/aws/en/sql/language-manual/sql-ref-syntax-qry-select-qualify
Official preparation steps
Follow the issuer's own sequence: review the exam guide, take the recommended training, register, run a system check, re-review for gaps, then sit the exam.
The certification page lays out the sequence explicitly: review the Data Engineer Professional Exam Guide, take the related training, register for the exam, check the technical requirements and run a system check, review the guide again to identify gaps, study to fill them, and take your exam. Registration happens through Databricks' exam delivery platform.
Recommended training includes the instructor-led Advanced Data Engineering With Databricks course and self-paced Databricks Academy courses covering declarative pipelines, data privacy, performance optimization, and automated deployment with bundles. The guide lists the current course titles.
For practice, the sample questions in the exam guide are the issuer-endorsed starting point. Free practice resources can supplement them, but the guide's outline is the authoritative list of what can be tested.
- Read the July 3, 2026 exam guide and its ten-section outline
- Take the recommended instructor-led or Academy training
- Register on the exam delivery platform and run the system check
- Re-review the guide about two weeks before your exam to catch updates
- Work the retired sample questions and map each to its outline section
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; Databricks Certified Data Engineer Professional | Databricks — Official certification page
Check your exam language and current instructions
The passing score is not stated in the retrieved official materials, and regional language pages currently disagree with the English guide on duration and recommended experience.
Neither the exam guide nor the certification page states a passing score or scaling method, so any raw mark you see discussed elsewhere is not something this guide can confirm. If a passing threshold matters to your planning, check Databricks' certification FAQ before your exam.
The Japanese and Korean regional certification pages retrieved for this guide list a 90-minute limit and a shorter recommended experience period, which conflicts with the English exam guide and English certification page. When sources disagree, the dated English exam guide, live as of July 3, 2026 and covering the currently live exam version, is the most current document; verify logistics on it and re-check close to your exam date.
Sources: Data Engineer Professional Exam Guide — Databricks exam guide PDF, live as of July 3, 2026; Databricks Certified Data Engineer Professional | Databricks — Official certification page; Databricks 認定データ エンジニア プロフェッショナル | Databricks — Japanese regional certification page; Databricks 공인 데이터 엔지니어 프로페셔널 | Databricks — Korean regional certification page
Sources
Facts checked against official Databricks documents: