Free study guide · Databricks
How to pass the Databricks Certified Data Analyst Associate
What the exam covers, how to plan your study, and how to read its questions. From the CertLumen course for this exam, kept up to date with the official exam guide.
- Questions
- 45
- Time
- 90 minutes
- Passing score
- Not published by Databricks; this course targets ≥ 80% on unseen mocks
- Exam fee
- USD 200 (plus applicable taxes)
- Valid for
- 2 years (recertify by taking the current live exam)
- Delivery
- Online proctored or test center
What the exam covers
A study plan that works
The default path for analysts who write SQL every day but haven't used every part of Databricks SQL, AI/BI dashboards and Genie. It covers all 21 lessons and 11 labs in course order, with a section drill at the end of each section and daily flashcards. Until you take the mocks, keep the practice setting that hides the mock-exam questions switched on, so the drills use only questions that aren't in a mock. Querying (n4, 20%), dashboards (n6, 16%) and analyzing queries (n5, 15%) get the most days because together they carry about half the exam. Mock Exam 1 and Mock Exam 2 are on days 20 and 21, and you sit the exam on day 22 or soon after. Plan on about 1–2 hours a day.
| Day | Focus | Time |
|---|---|---|
| 1 | Kick-off: the exam and the platform components | 70 min |
| 2 | Catalog Explorer: objects, access, certification and lineage | 95 min |
| 3 | Discover, certify, tag, trace and clean | 100 min |
| 4 | Cleaning lab and the ways data comes in | 105 min |
| 5 | Ingestion lab, SQL warehouses and Genie Code | 95 min |
| 6 | Aggregations, sorting and filtering | 75 min |
| 7 | Joins, set operations and federated queries | 100 min |
| 8 | Managed and external tables, CTAS and time travel | 75 min |
| 9 | Views, materialized views and streaming tables | 110 min |
| 10 | Photon, query history and caching | 70 min |
| 11 | Query insights, the query profile and liquid clustering | 75 min |
| 12 | Delta history for auditing, and fixing queries | 115 min |
| 13 | Building AI/BI dashboards and choosing charts | 70 min |
| 14 | Visualizations in notebooks and the SQL editor, and parameters | 115 min |
| 15 | Sharing, embedding, schedules and alerts | 115 min |
| 16 | Genie spaces (Genie Agents): purpose, components and setup | 80 min |
| 17 | Sharing, distributing and improving Genie spaces | 115 min |
| 18 | Data modeling, then the Unity Catalog namespace and roles | 130 min |
| 19 | Ownership, secure storage and PII protection | 110 min |
| 20 | Mock Exam 1 and review | 140 min |
| 21 | Mock Exam 2 and final review | 140 min |
Exam strategy
How the exam works
The Databricks Certified Data Analyst Associate exam checks that you can do everyday analytics on the Databricks Data Intelligence Platform: find and clean governed data, bring data in, write and tune SQL on SQL warehouses, build and share AI/BI dashboards and Genie spaces, model data and keep it secure. Most items are short scenarios that ask for the best option under a stated constraint, such as least privilege, no copies, always current or fastest for many users. Knowing that a feature exists isn't enough: you have to know when it's the right choice and what it doesn't do.
| Fact | Detail (from the exam page, the exam guide of Oct 30, 2025 and the certification FAQ) |
|---|---|
| Scored questions | 45 multiple-choice questions |
| Unscored items | May be included to gather statistics. They aren't identified and don't affect your score, and extra time is built in for them |
| Time limit | 90 minutes, about 2 minutes per question |
| Fee | USD 200 plus applicable taxes |
| Delivery | Online proctored or at a test center. You register on Webassessor, and Kryterion delivers the exam |
| Test aids | None allowed |
| Language | English |
| Prerequisites | None required. Related training and six months of hands-on data analysis on Databricks are highly recommended |
| Validity | 2 years. To recertify you take the full exam that is live at that time |
| Passing score | Not published. Databricks sets it statistically and says it can change |
| Recommended training | Data Analysis with Databricks (instructor-led or self-paced). The guide says the self-paced course will be replaced by two modules: AI/BI for Data Analysts and SQL Analytics on Databricks |
| Results and retakes | An unofficial result appears when you finish. A badge from Accredible follows after Databricks reviews the proctoring recording. You can retake after 14 days, and each attempt costs the full fee |
The exam page and the FAQ describe the format as multiple choice, but sample question 10 in the exam guide asks *Which two are features of the Query History?*, with B and D as the answer. So expect some items that name how many options to pick. This course has about one such *(Choose two.)* question in ten; they score only when both picks are right. Judge each option on its own as true or false before you look for the best ones.
You can't tell which items are unscored, so treat every question as if it counts. If one seems outside the outline, answer it quickly with your best guess and move on.
Where the questions are
The exam guide lists the nine sections without weights; the weights below come from the exam's web page. The question counts are this course's estimate on 45 scored items, and they are also how the course's two mock exams are built. Three sections, querying, dashboards and analyzing queries, carry 51% of the score.
| Section | Weight | ≈ Questions | Lessons | Highest-yield topics |
|---|---|---|---|---|
| n1 Understanding the Data Intelligence Platform | 11% | 5 | n1-l1, n1-l2 | Which component does what (Delta Lake, Unity Catalog, Data Intelligence Engine, Databricks SQL, Lakeflow, Mosaic AI), managed vs external on DROP, Catalog Explorer tabs, Marketplace |
| n2 Managing Data | 8% | 4 | n2-l1, n2-l2 | Certification as a governed tag, tagging and lineage in Catalog Explorer, NULLs and blanks, try_cast, deduplication with QUALIFY, constraints |
| n3 Importing Data | 5% | 2 | n3-l1 | S3 through external locations, Auto Loader vs COPY INTO vs read_files, OpenSharing (Delta Sharing), API patterns, UI upload limits |
| n4 Executing queries (SQL warehouses) | 20% | 9 | n4-l1 to n4-l5 | Warehouse types and sizing, Genie Code (Databricks Assistant), aggregates, joins and set operations, federation, managed vs external tables, time travel, MV vs streaming table vs dynamic view |
| n5 Analyzing Queries | 15% | 7 | n5-l1 to n5-l3 | Photon, Query History, result and disk caches, query profile and insights, liquid clustering, DESCRIBE HISTORY, fixing wrong queries |
| n6 Dashboards and Visualizations | 16% | 7 | n6-l1 to n6-l3 | Datasets, pages and filters, chart choice, parameters, publishing modes, sharing and embedding, schedules, alerts |
| n7 AI/BI Genie spaces (Genie Agents) | 12% | 5 | n7-l1, n7-l2 | Compute vs data credentials, instructions and trusted assets, permissions and distribution, Monitor tab, benchmarks, metadata refresh |
| n8 Data Modeling | 5% | 2 | n8-l1 | Star vs snowflake vs data vault, hubs, links and satellites, which model fits which medallion layer |
| n9 Securing Data | 8% | 4 | n9-l1, n9-l2 | Three-level namespace and volumes, grants and inheritance, roles and ownership, workspace object ACLs, masks, filters, ABAC |
| Total | 100% | 45 | 21 lessons |
People who write SQL every day usually do well on joins, filters and aggregates and lose points on platform behavior: which warehouse type a feature needs, which cache answered a query, what a dashboard viewer can see with shared data permissions, whose permissions a Genie space uses, and what certification or a tag does *not* do. Budget study time by weight and by unfamiliarity.
How to use this course
- Learn. Read the lesson, answer its self-checks and note any key point you would have missed.
- Lab. Do the lab on Databricks Free Edition. A few steps need a paid workspace (for example external tables on your own S3 bucket), and the lab says so. Seeing a result cache hit with no profile, or a dashboard viewer seeing data they have no grants on, makes the rule stick.
- Drill. After each section, take its drill untimed. Keep the practice setting that hides the mock-exam questions switched on until you've taken both mocks, so that the mocks stay unseen; the drills use all the other questions.
- Flashcards. Short daily sessions. Mark the cards you miss and repeat them. Defaults and limits decide many questions: 7 days, 30 days, 24 hours, 10 minutes, 45 minutes, 50 tables.
- Mocks. Two mock exams of 45 questions each, timed at 90 minutes and weighted like the real exam. They don't share questions. Take each in one sitting with no notes.
- Review loop. Keep an error log: the topic, why you missed it (knowledge gap, misread constraint, renamed product), the correct rule and where it lives in the course. Retest that topic two days later.
A repeated mock measures memory, not readiness. Count only first attempts. After both mocks, switch off the hide setting and use the full question bank for your weakest sections.
Readiness criteria
- ≥ 80% (36 of 45) on both mock exams, each taken timed in one sitting on the first attempt.
- No section below 70% across the two mocks combined. Small sections have only two questions per mock, so confirm them with their section drills.
- You finish a mock in ≤ 75 minutes and still have time to review flagged questions.
- Without notes, you can rebuild these cheat-sheet tables: warehouse types, the four caches, MV vs streaming table vs view, the two dashboard publishing modes, Genie permissions, and the Unity Catalog grants to read one table.
- You've done every lab at least once and can say what each one proved.
Databricks doesn't publish the passing score. Aiming for 80% on unseen mocks leaves a margin for a harder exam form, unfamiliar wording and exam-day nerves.
Time management
You have 90 minutes for 45 scored items plus any unscored ones, roughly 2 minutes per question. Recall items take 30–60 seconds, which buys time for code and fix-the-query items. Use three passes.
- First pass (about 55–60 minutes). Answer everything you can within about 90 seconds. If you're stuck between two options, pick the better one, flag the question and move on.
- Second pass (about 15–20 minutes). Return to flagged questions. Re-read the constraint words in the stem before you re-read the options.
- Final check (about 5 minutes). Make sure nothing is unanswered. Change an answer only for a concrete reason, such as a constraint you missed.
| Checkpoint | Target elapsed time |
|---|---|
| Question 15 | ≤ 25 minutes |
| Question 30 | ≤ 50 minutes |
| Question 45 | ≤ 70 minutes, then flagged questions only |
When time is short, answer every remaining question first, then go back to the flagged ones.
How to read the questions
A typical item gives a few sentences of context, some of it irrelevant, and then asks for the best choice. One or two constraint words usually decide the answer. Read in this order:
- Read the last sentence (the actual ask) first.
- Find the constraint words: no copies, always current, least privilege, viewers without access, each viewer sees only their rows, many concurrent users, one slow query, exactly once.
- Predict the answer before you look at the options.
- Eliminate every option that breaks a constraint, then choose from what's left.
| Constraint wording | What it usually rewards |
|---|---|
| "without copying or moving the data" | Lakehouse Federation for an external database, OpenSharing (Delta Sharing) for another organization, Marketplace for a third-party data product |
| "new files keep arriving", "only new files", "exactly once" | Auto Loader, which in SQL means a streaming table over STREAM read_files; COPY INTO for an idempotent scheduled SQL load |
| "always reflects the latest data", "no storage" | A standard view. A materialized view is only as fresh as its last refresh |
| "hundreds of users run the same heavy query" | A materialized view, refreshed on a schedule or with TRIGGER ON UPDATE |
| "queries wait in the queue" | More maximum clusters on the warehouse (scale out) |
| "one query spills to disk", "a single complex query is slow" | A larger cluster size (scale up), after fixing the query or the layout |
| "viewers don't have access to the tables" | Publish the dashboard with Share data permissions (the default) |
| "each viewer must see only their own rows" | Individual data permissions on a dashboard; a row filter or dynamic view in Unity Catalog. A Genie space always uses each user's permissions |
| "least privilege" | USE CATALOG + USE SCHEMA + SELECT for readers, CAN RUN or CAN VIEW on dashboards and Genie spaces, grants to groups |
| "approved calculation", "verified answer" | A Genie trusted asset: a parameterized example SQL query or a Unity Catalog SQL function |
| "prove accuracy improved" | Genie benchmarks, compared run to run |
| "a few percent of error is acceptable" | approx_count_distinct instead of count(DISTINCT ...) |
Words that point to the answer
Phrases in a question that usually point to one answer. The course has the full list.
| When the question says | Think of |
|---|---|
| the table no longer shows in SHOW TABLES, but its files are still in the bucket | It was an external table |
| a managed table was dropped by mistake three days ago | UNDROP TABLE |
| the user can find the table in Catalog Explorer but SELECT fails | Missing SELECT, USE SCHEMA or USE CATALOG |
| check mark next to the table, trusted by stewards | Certified: system.certification_status = certified |
| warn users off the old table, but existing jobs must keep running | Deprecate it (Assign certification > Deprecated) |
| a number looks wrong: which sources and jobs fed it? | Follow lineage upstream (Lineage tab > See lineage graph) |
| third-party or public data, no pipeline to build, always up to date | Databricks Marketplace: Get instant access |
| a partner on Databricks in another cloud, live read-only access, no exports | Databricks-to-Databricks OpenSharing (Delta Sharing) |
| new files land every few minutes, load each exactly once, schema may gain fields, SQL only | A streaming table over STREAM read_files (Auto Loader) |
| a scheduled SQL load that is safe to rerun | COPY INTO |
Revise from the cheat sheet
The course's cheat sheet has a page for each of these:
- Platform components, Catalog Explorer and Marketplace
- Discovery, certification, tags, lineage and cleaning
- Bringing data into Databricks
- Warehouses, Genie Code, SQL, tables and views
- Photon, caching, query profiles, clustering and history
- AI/BI dashboards, visualizations, parameters, sharing and alerts
- Genie spaces (Genie Agents)
- Data models and the medallion architecture
- Unity Catalog security
Practise with the real exam's mix
258 exam-style questions, 4 timed practice exams, flashcards and an AI tutor.