Free
- Untimed practice across all ten domains
- Every question explained against the objective it tests
- Domains by name and official weight
Databricks · Data Engineer Professional
Two options tell you to do the same thing. Only one of them says why correctly.
The Professional isn't a longer Associate. It takes the same platform — Auto Loader, declarative pipelines, Unity Catalog — and asks whether you'd still choose it under a cost ceiling, an SLA, or a retention policy. 59 questions in 120 minutes across ten domains, one year of hands-on work recommended, no test aids of any kind, and a pass mark Databricks doesn't publish.
- Written to the exam guide version live since 3 July 2026, not the six-domain outline it replaced
- Weighted to the official split, where 65% of the exam sits in domains worth 10% or less
- Python-first, the way this exam is — the Associate is the SQL-first one
Free practice across all ten domains. No card required.
Data Engineer ProfessionalDeveloping Code for Data Processing using Python and SQLHard
A gold-layer object serves an hourly dashboard. Its silver source is not append-only: rows are updated in place when late corrections arrive. The team wants the gold object refreshed without recomputing the full aggregation every hour. Which choice is correct, and for the right reason?
Select one.
One choice only. Your earliest pick was dropped.
Where this question sits in the exam
Developing Code for Data Processing using Python and SQL is 22% of the Data Engineer Professional exam. Our mocks use the same mix.
0 of 1 selected
An original item written against the Databricks exam guide, last checked 18 September 2026. Not a real exam question.
How it's scored
The pass mark isn't just unpublished. It moves.
Databricks sets passing scores by statistical analysis and revises them as questions are swapped into the exam. Their own certification FAQ gives that as the reason they don't publish the number: because it changes, they don't print it.
The 70% you'll find quoted for this exam is not a Databricks number. Neither is the 60-question count that still appears on prep sites — that was true of the guide retired in 2026, and the current form has 59 scored questions.
We still have to put a line somewhere. A readiness read means nothing without one. So we assume 80% weighted accuracy.
We set it deliberately high. If we're too strict, you practise a week longer than you needed to. If we're too lenient, you pay USD 200 to find out. Those two mistakes don't cost the same, so we err toward the one that costs you time.
Data Engineer Professional59 questions · 120 min
Accuracy by domain, weighted as the exam weights them · assumed 80% — ours, not Databricks'
Code (Python & SQL)22%78%
Ingestion7%72%
Transformation10%69%
Sharing5%81%
Monitoring10%64%
Cost & Performance13%Furthest behind51%
Security10%67%
Governance7%74%
Debugging10%70%
Modeling6%76%
Cost & Performance Optimization is 13% of the form and your weakest domain. That's where the next hour goes.
Recertifying here means the full current exam
The credential lasts two years. Databricks publishes a 25-question renewal exam for the Data Engineer Associate and none for the Professional. The guide and the certification page agree: recertifying means sitting whatever version of the full exam is live then.
On a platform shipping this fast the guide will have moved. Features that did not exist when you certified are on the form, and topics you revised for may be gone.
What the questions look like
The hard part isn't picking the feature. It's picking the reason.
Two options, one action, two reasons
The guide's own published samples include items where more than one option proposes the same change and they differ only in the justification behind it. Eliminating the obviously wrong feature gets you to two options and no further; the rest of the work is deciding which explanation actually holds.
Constraints from production, not from a tutorial
Stems carry service-level agreements, cost ceilings, retention policies and compliance requirements. The same feature is the right answer in one stem and the wrong one in the next, and only the constraint tells you which.
Python and SQL, from memory
Code examples are primarily in Python and SQL, and the largest domain on the exam has Python in its name. No test aids are permitted — the guide names API documentation specifically — so configuration keys, method names and CLI syntax come from recall.
The loop
Practice, sit, review, repeat
Step 1 is free. Steps 2 to 4 need a subscription.
certs.acedify.com/data-engineer-professional
Practice · untimed
HardA gold-layer object serves an hourly dashboard. Its silver source is not append-only: rows are updated in place when late corrections arrive. The team wants the gold object refreshed without recomputing the full aggregation every hour.
Which choice is correct, and for the right reason?
A streaming table, because it processes only new rows and so costs less than recomputing
A materialized view, because a materialized view always refreshes incrementally and never fully recomputes
A materialized view, because a streaming table's source must be append-only and an in-place update would break the stream
A streaming table with skipChangeCommits enabled, so that updates in the source are ignored
The source receives in-place updates, so it is not append-only. That rules out a streaming table before cost enters the picture. A materialized view is the object that can still refresh; claiming it always refreshes incrementally is the same action with the wrong reason. Domain: Developing Code for Data Processing using Python and SQL.
Click a step to see it.
What you get back
No official line. So we show you ours.
Your readiness is a range, because it's an inference from the questions you've answered rather than a result you've earned. For a certification the outcome is binary, so the question is whether the whole range sits above the line.
The difference here is whose line it is. AWS publishes 720 and we measure you against their number. Databricks publishes nothing, so we measure you against ours — 80% weighted accuracy, set high on purpose, drawn dashed and labelled everywhere it appears.
Ten domains makes this harder to eyeball than it looks. A bad score in Data Sharing costs you about three questions; the same score in the code domain costs you thirteen.
When the range stops narrowing and stops rising, more drilling isn't buying you anything, and that's worth knowing before you book.
If the real pass mark is below ours, the worst case is that you sat more practice than you needed. We'd rather make that mistake than the other one.
Data Engineer Professional · 4 mocks completed
Where you stand
Borderline
70–90
assumed 80% — ours, not Databricks'
Your range crosses our assumed line. On today's evidence it's a coin flip, and we'll say so.
Three states
Try the three states. We won't show you a percentage chance of passing — nobody can calculate one honestly without years of real outcomes to calibrate against.
What we don't do
We won't show you a percentage chance of passing. Nobody can calculate one honestly without years of real outcomes behind it.
What it actually tests
Ten domains, and there's nothing safe to skip
On the Associate, two domains are 43% of the exam and you can plan around that. Here the biggest is 22%, the second is 13%, and everything else is 10% or under — which adds up to 65% of the form sitting in the small domains. There's no pair you can master to carry you, and no domain small enough to write off: the smallest is still around three questions on a form where nobody will tell you how many you need.
| Domain | Weight | What it covers |
|---|---|---|
| Developing Code for Data Processing using Python and SQL | 22% · about 13 questions | The biggest block by some distance. Project structure for Declarative Automation Bundles, third-party library and dependency management, Pandas and Python UDFs, AUTO CDC, streaming tables versus materialized views, and unit tests with assertDataFrameEqual and assertSchemaEqual. |
| Data Ingestion & Acquisition | 7% · about 4 questions | Ingesting Delta, Parquet, ORC, AVRO, JSON, CSV, XML, text and binary from message buses and cloud storage, and append-only pipelines handling batch and streaming together. |
| Data Transformation, Cleansing, and Quality | 10% · about 6 questions | Window functions, joins and aggregations at scale in Spark SQL and PySpark, plus quarantining bad data in declarative pipelines or Auto Loader. |
| Data Sharing and Federation | 5% · about 3 questions | The smallest domain — roughly three questions. Delta Sharing between Databricks deployments and to external platforms, and Lakehouse Federation with governance across supported sources. |
| Monitoring and Alerting | 10% · about 6 questions | System tables for cost, utilisation and audit; Query Profiler and Spark UI; REST API and CLI monitoring; pipeline event logs; SQL Alerts and job notifications. |
| Cost & Performance Optimization | 13% · about 8 questions | Second-biggest. Managed tables and operational overhead, deletion vectors, liquid clustering, data skipping and file pruning, Change Data Feed, and reading the query profile for bad skipping and bad joins. |
| Ensuring Data Security and Compliance | 10% · about 6 questions | Workspace object ACLs and least privilege, row filters and column masks, hashing, tokenisation, suppression and generalisation, PII masking in batch and streaming, and purging against retention policy. |
| Data Governance | 7% · about 4 questions | Descriptions and metadata that make enterprise data discoverable, and the Unity Catalog permission inheritance model. |
| Debugging and Deploying | 10% · about 6 questions | Spark UI, cluster logs, system tables and query profiles for diagnosis; job repairs and parameter overrides; Declarative Automation Bundles and Git Folders in CI/CD. |
| Data Modeling | 6% · about 4 questions | Scalable Delta models, liquid clustering versus partitioning and ZORDER, and dimensional models for analytical workloads. |
Domain weights are published on the Databricks certification page. The exam guide lists the same ten sections without percentages. Last checked 18 September 2026.
Weights are published on the certification page only. The exam guide lists the same ten sections without percentages.
Exam facts
Null fields are omitted. We do not print TBA.
- Credential
- Databricks Certified Data Engineer Professional
- Level
- Professional
- Current version
- Live since 3 July 2026
- Length
- 120 minutes
- Scored questions
- 59
- Unscored questions
- Present, count not published
- Question types
- Multiple choice
- Pass mark
- Not published
- Domains
- 10
- Fee
- USD 200 + local tax
- Delivery
- Online or test centre, proctored
- Test aids
- None, including API documentation
- Languages
- English, 日本語, Português BR, 한국어
- Prerequisite
- None
- Recommended experience
- One year of hands-on data engineering on Databricks
- Validity
- 2 years
- Retake wait
- 14 days between attempts, unlimited attempts, full fee each
- Recertification
- Full current exam — no renewal exam for this credential
Pricing
Practice free. Pay when you want to know where you stand.
Pro
- Timed mocks at 59 questions in 120 minutes
- Weighted accuracy across every mock you've sat
- Your readiness range, and how it moves
Everything on every exam, one subscription
One subscription covers the Databricks certifications, AWS, the Microsoft certifications, the Salesforce certifications, the Claude certifications, TOEFL, IELTS and the SAT.
The exam fee is separate and goes to Databricks. We're prep, not registration.
Start practising for freeFAQ
Data Engineer Professional questions
Ten domains. Find out which one is furthest behind.
Free practice across all ten, no card. The timed mocks and the readiness range are there when you want them.
Start practising for freeAlso available for the Data Engineer Associate, Data Analyst Associate, Machine Learning Associate and the rest of the Databricks track on the same subscription.