What changes on May 4, 2026, what stays the same, and how we are preparing you for it without the stress.
A note from Team BricksNotes.
Databricks has published a new version of the Data Engineer Associate exam guide. It goes live on May 4, 2026. If your exam is on or after that date, this is the version you should prepare against.
We rebuilt our practice exam to follow it on the same day, so you can keep learning without worrying about which guide is current. That is the whole reason we exist. We do the busy work of tracking changes, so you can focus on understanding the system.
Use the new guide if your exam is on or after May 4, 2026. If your exam is before that date, the older guide still applies. Most learners reading this today will sit the new version, so the rest of this article is about that one.
Nothing dramatic has changed in the format. The shift is in what is tested and in the product names Databricks now uses.
The exam checks whether you can use the Databricks Data Intelligence Platform to do foundational data engineering work. That means understanding the platform and its workspace, ingesting and loading data, transforming and modeling it with PySpark and SQL, orchestrating with Lakeflow Jobs, working with CI/CD, and handling troubleshooting, monitoring, optimization, governance and security.
In plain words, it asks: can you build and run a reliable pipeline on Databricks, and can you reason about it when something goes wrong?
The core building blocks. Architecture, Delta Lake, and Unity Catalog. You should also know the compute services, what each one is good at, what its limits are, what it costs, and which one to pick for a given workload. This now includes Lakebase, a managed Postgres service in the lakehouse that is generally available.
This area is broader than before. Expect questions on:
COPY INTO for incremental loads from cloud object storage (ADLS, S3, GCS) into Unity Catalog tablesReading bronze tables, cleaning nulls, standardizing types, and writing silver tables in PySpark or SQL. Combining DataFrames with inner, left, broadcast, multi-key, cross joins, plus union and union all. Manipulating columns and rows, exploding arrays, deduplicating, and aggregating with count, approximate count distinct, mean, and summary.
You should also know the basic Spark tuning knobs, such as spark.sql.shuffle.partitions, spark.default.parallelism, executor and driver memory, and spark.sql.autoBroadcastJoinThreshold, and how to remeasure performance after changing them.
For the gold layer, know the difference between materialized views, views, streaming tables, and tables in Unity Catalog, and how to apply data quality checks on silver and gold datasets.
Control flow with retries, branching, and looping. Configuring notebook, SQL query, dashboard, and pipeline tasks, and wiring their dependencies in the DAG-based task graph. Scheduling jobs, and choosing between time-based triggers and data-driven ones like file arrival or table updates, based on data availability.
This area has the largest naming change. Be ready for:
Reading the Lakeflow Jobs run history to spot trends, interpreting DAG task graphs to find upstream blockers, and tracking run times and failure rates. Spotting common bottlenecks like data skew, shuffling, and disk spilling in the Spark UI. Knowing what Liquid Clustering and predictive optimization do. Diagnosing cluster startup failures, library conflicts, and out-of-memory issues.
For governance, know the difference between managed and external tables in Unity Catalog and how to convert between them. Apply GRANT, REVOKE, and DENY to users, groups, and service principals at the right level of the hierarchy. Understand column-level masking and row-level security, and how Unity Catalog ABAC policies centrally control row filtering and column masking for sensitive data. Note that Delta Sharing has evolved into OpenSharing, an open protocol for sharing data and AI assets.
A few themes stand out compared to older guides:
If you have been preparing against an older guide, the concepts are still valid. You mostly need to update the names you reach for, and add a little time on ABAC, Liquid Clustering, and Automation Bundles.
For completeness, the official guide points to:
Our practice exam already follows the May 2026 blueprint. Concretely:
It is free to try with a single email. No account, no spam.
We also keep our lessons aligned with the same blueprint, so you can start from the basics and walk forward at your own pace.
We started BricksNotes because we remembered what it felt like to prepare for a certification while the product underneath kept changing names. It is tiring. It pulls energy away from the part that matters, which is understanding how the system actually behaves.
So we made a small promise to ourselves and to you. Whenever Databricks updates the exam guide, we will update our practice exam, our lessons, and our writing in step. You should not have to chase any of it.
Show up, practice, ask "why" until it clicks. We will carry the rest, so you can learn and grow without stress.
— Team BricksNotes
BricksNotes is an independent learning platform and is not affiliated with or endorsed by Databricks. Databricks is a registered trademark of Databricks, Inc.