Databricks Data Engineer Associate: what changed in the May 2026 guide
Databricks replaced the exam outline twice in about ten months. The version live since May 4, 2026 has seven weighted sections, a long list of new topics, several dropped ones and a new name for almost every product. Here is every change, sourced from the exam guide, the certification FAQ and Databricks documentation.
The short version
The exam guide dated May 2026 says it covers the exam version as of May 4, 2026. The guide before it covered the exam as of July 25, 2025. Those two documents describe different exams. The July 2025 outline had five sections with no published weights and leaned on Delta Sharing, Lakehouse Federation, lineage, audit logs and Databricks Connect. The May 2026 outline has seven sections with weights, 33 bullets in total, and is built around ingestion with Lakeflow Connect, transformation in PySpark and SQL, orchestration with Lakeflow Jobs, CI/CD with Git folders and bundles, troubleshooting in the Spark UI, and governance in Unity Catalog, including ABAC.
The format did not move: 45 scored multiple-choice questions, 90 minutes, no test aids, online or at a test center, two-year validity. Databricks still does not publish a passing score. What changed is what you are asked, and the names of the products you are asked about.
Two rules from Databricks' certification FAQ matter here: "there will only ever be one version of each exam", so there is no overlap period, and the exam guide "is the source of truth for what is on a currently live exam". If a study resource and the guide disagree, the guide wins.
Change 1: seven weighted sections
The old guide listed five sections without weights. The new one publishes a weight for each of seven.
| # | May 2026 section | Weight | Bullets | Closest July 2025 section |
|---|---|---|---|---|
| 1 | Databricks Intelligence Platform | 6% | 2 | Databricks Intelligence Platform |
| 2 | Data Ingestion and Loading | 21% | 7 | Development and Ingestion |
| 3 | Data Transformation and Modeling | 22% | 7 | Data Processing & Transformations |
| 4 | Working with Lakeflow Jobs | 16% | 4 | Productionizing Data Pipelines (part) |
| 5 | Implementing CI/CD | 10% | 4 | Productionizing Data Pipelines (part) |
| 6 | Troubleshooting, Monitoring, and Optimization | 10% | 5 | Productionizing Data Pipelines (Spark UI) and Development (debugging) |
| 7 | Governance and Security | 15% | 4 | Data Governance & Quality |
Change 2: what is new in the outline
Comparing the two guides line by line, these topics appear in the May 2026 outline and did not appear in July 2025:
- Ingestion: COPY INTO; Auto Loader's directory listing and file notification modes; Lakeflow Connect standard and managed connectors; JDBC/ODBC and REST clients; choosing an ingestion method; JSON and nested data.
- Transformation: joins, union and union all; column and row operations including explode; deduplication and aggregates; the tuning parameters spark.sql.shuffle.partitions, spark.default.parallelism, executor and driver memory and spark.sql.autoBroadcastJoinThreshold; Gold-layer materialized views, views, streaming tables and tables; data quality checks.
- Orchestration: retries, branching and looping; task types in a DAG; scheduled, file arrival and table update triggers, and choosing between them.
- CI/CD: branches, commits and pull requests from Git folders; bundle variables and targets; the Databricks CLI to validate, deploy and manage bundles.
- Operations: run history trends, job statuses, skew, shuffle and spill in the Spark UI, Liquid Clustering, predictive optimization, startup failures, library conflicts and out-of-memory errors.
- Governance: converting managed and external tables; GRANT, REVOKE and DENY; column masks and row filters; Unity Catalog ABAC policies.
Change 3: what was dropped
The July 2025 outline asked you to use Delta Sharing, explain its advantages, limits, types and cross-cloud cost, use lineage in Unity Catalog, identify how audit logs are stored, use Lakehouse Federation, and use Databricks Connect in a workflow. None of those bullets appears in the May 2026 outline. The current guide even labels its sample question about audit logs as retired from a previous version. If your notes have a chapter on Delta Sharing, you can archive it.
One caution: the guide says "additional related topics" can still appear, so dropped means unlisted, not forbidden.
Change 4: the product renames
The exam uses the current names. Older courses and books use the old ones. This table maps them, with the Databricks source for each.
| Current name (what the exam says) | Previous name | Where Databricks says so |
|---|---|---|
| Lakeflow Jobs | Databricks Workflows / Jobs | May 2026 guide, Section 4; jobs documentation |
| Lakeflow Spark Declarative Pipelines (docs: "Lakeflow pipelines") | Delta Live Tables (DLT), then Lakeflow Declarative Pipelines | Guide bullet 5.3; docs page "Where is DLT?" says the product formerly known as DLT has been updated to Lakeflow pipelines. In Python, from pyspark import pipelines as dp replaces import dlt, which still works |
| Declarative Automation Bundles (the guide also writes "Automation Bundle") | Databricks Asset Bundles (DABs) | Guide bullets 5.2 to 5.4; bundles release notes of March 16, 2026. The FAQ says the rename is non-breaking: the bundle CLI command and existing configuration do not change |
| Git folders | Databricks Repos | Guide bullet 5.1 ("formerly Databricks Repos"); repos documentation |
| Lakeflow Connect (managed and standard connectors) | New product area | Guide Section 2; Lakeflow Connect documentation |
| Standard access mode | Shared access mode | Compute documentation: "standard access mode (formerly shared access mode)" |
| Dedicated access mode | Single user access mode | Compute documentation: dedicated is the latest version of single-user access mode |
"High-concurrency cluster", used in one of the guide's retired sample questions, is a legacy term. Expect the current access-mode names in new questions.
Change 5: details that are easy to get wrong
How to adjust your preparation
- 1
Check the guide first
Open the exam guide linked from the Databricks certification page and confirm its 'exam version as of' date. It changed in July 2025 and May 2026 and can change again; the May guide no longer promises a two-week notice.
- 2
Relearn the names
Lakeflow Jobs, Lakeflow Spark Declarative Pipelines, Declarative Automation Bundles, Git folders. Answer options are written in the new vocabulary.
- 3
Spend 43% of your time on Sections 2 and 3
Build one pipeline end to end: COPY INTO or Auto Loader into bronze, a cleaned silver table with joins and deduplication, and a Gold materialized view.
- 4
Add the operations layer
Schedule it as a Lakeflow Job with a file-arrival trigger, deploy it with a bundle to a second target, and read its run history and a Spark UI stage page. Sections 4 to 6 are 36% of the exam.
- 5
Practice on questions written for May 2026
Older banks still ask about Delta Sharing and still say Workflows and DLT. Use tests weighted 6/21/22/16/10/10/15 and read every option's explanation.
Start with the 10 free Databricks Data Engineer Associate questions, then compare with the AWS SAA-C03 study plan if cloud architecture is your next step. All courses are on the catalog.
Practice on the May 2026 guide, not the old one
Six 45-question practice tests weighted 6/21/22/16/10/10/15, written with the current product names, every option explained.
Databricks exam changes FAQ
When did the Databricks Data Engineer Associate exam change?
The current exam guide says it covers the exam version as of May 4, 2026. The previous guide covered the exam as of July 25, 2025. Databricks says only one version of each exam is ever live, so from May 4, 2026 the May 2026 guide is the one to study.
What is new in the May 2026 exam guide?
A rewrite into seven weighted sections. New topics include COPY INTO, Lakeflow Connect standard and managed connectors, JDBC/ODBC and REST ingestion, joins, unions, explode and deduplication in PySpark, Spark tuning parameters, Gold-layer objects, job trigger types, Git folders, Declarative Automation Bundles with the CLI, the Spark UI, Liquid Clustering, predictive optimization, row filters, column masks and ABAC.
What was removed?
Delta Sharing, Lakehouse Federation, lineage and audit logs no longer appear in the outline, and neither does Databricks Connect. The guide's sample question on audit logs is marked as retired from a previous version.
Did the format change?
No. The exam is still 45 scored multiple-choice questions in 90 minutes, online or at a test center, with possible unscored items. Databricks still does not publish a passing score.
Which product names does the exam use?
The current ones: Lakeflow Jobs (formerly Workflows), Lakeflow Spark Declarative Pipelines (formerly Delta Live Tables), Declarative Automation Bundles (formerly Databricks Asset Bundles), Git folders (formerly Repos), Lakeflow Connect, and standard and dedicated access modes (formerly shared and single user).
More guides: Free Databricks Data Engineer Associate questions · AWS SAA-C03 study plan · Free AWS SAA-C03 questions · all guides