The Databricks Certified Data Engineer Associate examination, in the version of May 4, 2026, presents 45 scored multiple-choice questions in 90 minutes across seven sections. Data ingestion and loading and data transformation and modelling together carry 43 percent of the paper, and the examination presupposes PySpark ETL work on the Databricks Data Intelligence Platform. The fee is 200 dollars, no prerequisite is required, and certification runs two years.
This program is part of the expanding PassProfessor catalog — join the notify list and we’ll email you when enrollment opens.
What this credential certifies
This examination certifies a working data engineer at the platform’s bench. Databricks describes its audience plainly: individuals who use the Data Intelligence Platform to execute foundational data engineering tasks — ingesting data, loading it, transforming and modelling it with PySpark, orchestrating it with Lakeflow Jobs, shipping it under CI/CD discipline, and keeping it governed in Unity Catalog. The Associate level is the craft level: not architecture in the abstract, but pipelines that run, land clean data, and can be trusted by the analysts downstream.
The current exam guide, covering the version live as of May 4, 2026, distributes 45 scored questions across seven sections — and its centre of gravity is unmistakable. Ingestion and loading at 21 percent and transformation and modelling at 22 percent together make 43 percent of the paper; add Lakeflow Jobs at 16 and governance and security at 15, and three quarters of the examination is the pipeline lifecycle itself. The platform section, by contrast, is a deliberate 6 percent: candidates are expected to arrive knowing what the platform is, and to be examined on what they can do with it.
The guide also records the platform’s recent renaming wave, and the examination uses the current names: Lakeflow Jobs for orchestration, Lakeflow Connect for ingestion connectors, Lakeflow Spark Declarative Pipelines, Git Folders in place of the former Repos, and Declarative Automation Bundles in place of the former asset bundles. Preparation material that still teaches the old vocabulary dates itself — and risks the candidate, since the questions are written in the platform’s present tense.
The examination at a glance
| Item | Verified position |
|---|---|
| Certification | Databricks Certified Data Engineer Associate, awarded by Databricks |
| Current version | Exam guide covering the exam version as of May 4, 2026; Databricks advises checking the live guide again about two weeks before the exam date |
| Examination | 45 scored multiple-choice questions in 90 minutes; the examination may also include unscored items, which are not identified and do not affect the score, with additional time factored in for them |
| Passing standard | Databricks does not publish a numeric passing score in the exam guide, and none is stated here; candidates prepare to the full outline rather than to a quoted threshold |
| Prerequisites | None required. Course attendance and six months of hands-on Databricks experience are highly recommended by the guide |
| Delivery | Online proctored or at a test centre; no test aides are allowed |
| Examination fee | Databricks fee: USD 200 plus applicable taxes — a third-party fee charged by Databricks for its own examination |
| Certification period | Two years. Recertification is required and is earned by taking the current live examination again |
| Independence | PassProfessor is independent and is not affiliated with Databricks. Databricks, Delta Lake, and Unity Catalog are Databricks’ technologies; the certification is Databricks’ |
What the examination covers
Databricks Intelligence Platform — 6%
The foundation section: the platform’s core components and architecture, Delta Lake and Unity Catalog, and the compute services — their characteristics, limitations, and cost models — with the judgement to select the suitable option for each workload.
Data Ingestion and Loading — 21%
Bringing data in: batch, streaming, and incremental patterns; local files and Lakeflow Connect’s standard and managed connectors; COPY INTO loading files incrementally from cloud object storage into Unity-Catalog-governed tables; Auto Loader with schema enforcement and evolution; JDBC, ODBC, and REST landing paths; choosing among ingestion methods by volume, frequency, type, and governance; and landing semi-structured and unstructured data such as nested JSON into Delta tables.
Data Transformation and Modeling — 22%
The examination’s largest section: cleaning bronze tables with PySpark and SQL into silver — nulls handled, types standardised; combining DataFrames across the join family, unions, and column and row manipulation including exploding arrays; deduplication and aggregation; basic Spark tuning parameters and re-measurement; building gold-layer objects — materialized views, views, streaming tables, and tables for BI and analytics in Unity Catalog; and applying data-quality checks so silver and gold datasets are reliable.
Working with Lakeflow Jobs — 16%
Orchestration: control flow with retries and conditional tasks — branching and looping; configuring notebook, SQL query, dashboard, and pipeline tasks and their dependencies in the DAG-based task graph; and scheduling with the trigger family — scheduled, file-arrival, and table-update — choosing between time-based and data-driven triggers by data availability and pipeline dependency.
Implementing CI/CD — 10%
Engineering discipline on the platform: the Git Folders workflow of branches, commits, pushes, and pull requests inside the workspace; environment-specific configuration with Declarative Automation Bundle variables and overrides promoting one codebase across development, test, and production targets; deploying bundles that package jobs, declarative pipelines, and other assets; and driving it all through the Databricks CLI in automated workflows.
Troubleshooting, Monitoring, and Optimization — 10%
Running pipelines in anger: reading job performance trends from run history against baselines; monitoring pipeline health in the Jobs UI through statuses, DAG graphs, run times, and failure rates; identifying skew, shuffling, and spilling from stage-level Spark UI metrics; Liquid Clustering and predictive optimization; and diagnosing failures of clusters to start, library conflicts, and out-of-memory errors.
Governance and Security — 15%
Unity Catalog as the governing layer: managed against external tables and the operations on each; access control through GRANT, REVOKE, and DENY applied to users, groups, and service principals at the right level of the hierarchy; column-level masking and row-level security restricting visibility by group; and attribute-based access control policies centralising row filtering and column masking for sensitive data.
Who this certification serves
The Associate examination serves practising and aspiring data engineers whose work runs on Databricks: pipeline builders, analytics engineers moving toward the engineering side, platform-adjacent developers taking on ETL responsibility, and engineers migrating from other Spark environments to the Data Intelligence Platform. Databricks requires no prerequisite but highly recommends six months of hands-on experience, and the outline earns that recommendation — its items presuppose PySpark written, jobs run, and tables governed, not merely concepts recognised. It is the foundation on which Databricks’ professional-level engineering certification builds.
The route, in order
Step 1 — Build the recommended base. No prerequisite gates the examination, but the guide highly recommends course attendance and six months of hands-on Databricks experience. Candidates short of that base should run real pipelines in a workspace — ingestion, transformation, a scheduled job — before treating the outline as a checklist.
Step 2 — Study to the May 2026 guide. Take the seven-section outline in the exam guide for the version of May 4, 2026, as the syllabus of record, in the platform’s current vocabulary. Databricks advises checking the live guide again about two weeks before the exam date, because the guide is updated whenever the examination changes.
Step 3 — Register and sit the examination. Register through Databricks’ certification system. The fee is USD 200 plus applicable taxes, charged by Databricks. The sitting is online proctored or at a test centre, with no test aides allowed: 45 scored multiple-choice questions in 90 minutes, with any unscored items unidentified and additional time factored in for them.
Step 4 — Recertify on the two-year cycle. Certification is valid for two years. Recertification is required to maintain it and is earned by taking the current live examination — the whole paper again, in its then-current version — so the holder’s calendar should carry the expiry from the day the credential issues.
What makes this examination demanding
Forty-five questions in 90 minutes is two minutes a question, and the questions are scenario-dense: a pipeline described, a symptom reported, four interventions offered, of which the managed, governed, current-platform answer must be separated from three plausible older or manual ones. Code reading is continuous in the transformation section. The renaming wave adds a quiet tax — candidates fluent in the platform’s former vocabulary must translate under pressure — and the spread across seven sections means a whole discipline such as CI/CD, a tenth of the paper, can be lost by a candidate who never promoted code formally.
The credential at work
A certified Associate’s attestation is pipeline trust: bronze lands complete, silver is clean and deduplicated, gold is modelled for the analysts who consume it, jobs retry and branch as designed, code promotes from development to production without hand edits, and access to every table can be accounted for in Unity Catalog. Employers read the credential as the platform vendor’s own statement that the holder works the current platform in its current idiom. The two-year term, renewed by passing the live examination again, keeps that statement in the platform’s present tense as deliberately as the questions are.
A closer look
The Associate paper’s most useful self-description is hidden in its weights: it is an examination about data in motion under governance. Ingestion moves data onto the platform, transformation moves it toward trustworthiness, Jobs move it on schedule, CI/CD moves the code that moves it, and Unity Catalog governs every movement. Candidates who organise their preparation around that single sentence find the seven sections stop competing for attention and start cooperating — a governance question becomes recognisable as a transformation question viewed from the access side, and a troubleshooting question as an orchestration question viewed after failure. The platform section’s small weight then reads correctly: motion needs only a compact map of the territory it crosses, but it needs that map exact.
What the PassProfessor program will cover
Practice in the examination’s own shape. The Databricks Certified Data Engineer Associate examination is a proctored, computer-delivered examination, and the program in the works practises in that shape throughout: timed sets at the examination’s pace, items in its formats, and domain sets matched to the published outline’s weights. Candidates should arrive having already made their pacing mistakes in practice.
Written from the primary sources. The program is written from Databricks’s own published outline and programme documents — the same documents the examination is built from. Where this credential’s rules differ from its neighbours’, and in this family they differ in eligibility, weighting, and maintenance, the program teaches this credential’s rule and names the difference rather than teaching a family average.
The whole candidate, not only the test-taker. Around the item practice, the program carries the certification route as Databricks publishes it — eligibility, application, the sitting itself, and the maintenance cycle — because an examination is a means to a credential, and candidates plan better when the destination is mapped.
Preparation only. This is preparation for an examination administered by Databricks. PassProfessor is not affiliated with Databricks, no completion of this program is reported to Databricks, and nothing in the program substitutes for any step in the certification process.
Tell me when enrollment opens. The Databricks Certified Data Engineer Associate program is in the works and enrollment is not yet open. Leave an email address and PassProfessor will write to you when the program is ready — and only about this program. No date is promised, because the program will open when it meets the standard described above, not before.
Frequently asked questions
How many questions are on the Databricks Data Engineer Associate examination?
45 scored multiple-choice questions in 90 minutes. The examination may include additional unscored items used to gather statistical information; they are not identified, do not affect the score, and additional time is factored in for them.
What is the passing score for this examination?
Databricks does not publish a numeric passing score in the exam guide, so none is stated here. Candidates should prepare to the full seven-section outline rather than to a quoted threshold, and treat any third-party percentage claim with caution.
What are the seven exam sections and weights?
Databricks Intelligence Platform 6 percent, Data Ingestion and Loading 21 percent, Data Transformation and Modeling 22 percent, Working with Lakeflow Jobs 16 percent, Implementing CI/CD 10 percent, Troubleshooting, Monitoring, and Optimization 10 percent, and Governance and Security 15 percent — the exam guide for the version of May 4, 2026.
Are there prerequisites for the Databricks Data Engineer Associate?
None required. The guide highly recommends course attendance and six months of hands-on Databricks experience, and its items presuppose that working familiarity in practice.
What does Databricks charge for the examination?
USD 200 plus applicable taxes — Databricks’ fee for its own examination, not a price for anything PassProfessor sells.
How long is the certification valid?
Two years. Recertification is required to maintain the certification and is earned by taking the current live examination again.
Where is the examination taken?
Online under proctoring or at a test centre, with no test aides allowed, under the rules of Databricks’ certification programme.
What products does the examination name — Lakeflow Jobs or Workflows?
The current names. The May 2026 guide examines Lakeflow Jobs, Lakeflow Connect, Lakeflow Spark Declarative Pipelines, Git Folders, and Declarative Automation Bundles, noting former names where Databricks itself notes them. Preparation should use the current vocabulary.
Does the examination require writing PySpark?
It examines the ability to perform ETL tasks using PySpark and to reason about PySpark and SQL transformation code — reading, completing, and judging it inside scenarios — within a multiple-choice format. Hands-on PySpark practice is the preparation the outline presupposes.
What will the PassProfessor Databricks Data Engineer Associate program cover?
Preparation for the examination as Databricks publishes it: scenario and code-reading practice at the examination’s pace, organised by the seven sections of the May 2026 guide at their published weights, in the platform’s current vocabulary, plus the certification and recertification route as Databricks states it. The program is in the works and enrollment is not yet open.