Spark Developer Associate resources

This is the most code-centric of the Databricks exams. DataFrame API work carries 30%, architecture 20% and Spark SQL 20%. The fastest preparation is writing Spark and reading query plans, not watching someone else do it. The official Spark documentation is excellent and free.

Start with the official material

The exam guide and its weights: DataFrame and DataSet API applications 30%, architecture and components 20%, Spark SQL 20%, troubleshooting and tuning 10%, structured streaming 10%, Spark Connect 5%, pandas API on Spark 5%.

The Apache Spark documentation. Genuinely one of the better official references, and free. The programming guides cover most of the API ground.

Databricks Academy’s Spark material for the platform-specific framing.

Write code and read plans

• Learn which operations are actions. Write a chain of transformations, confirm nothing runs, then add count() and watch execution start. The lazy evaluation question is free marks once this is intuitive.
• Read an execution plan before and after a filter, and find the predicate pushdown.
• Force and prevent a broadcast join, and see the shuffle appear and disappear.
• Create skew deliberately — a key with far more rows than the rest — and watch three tasks dominate a stage.
• Cause a spill by leaving shuffle partitions badly sized, then tune them so partitions fit in memory.
• Replace a row-at-a-time Python UDF with built-in expressions and measure the difference.
• Run a structured streaming query with a checkpoint, kill it, restart it, and confirm it resumes rather than reprocessing.
• Add a watermark and observe old state being dropped.

Each maps to a recurring exam theme, and each is faster to learn by doing than by reading.

The passing score problem

Databricks publishes no passing score. Any specific figure quoted elsewhere is a guess. A weighted practice test is the only calibration available — the free 20-question test here matches the published sections.

Worth paying for

A practice test with explanations.

Compute for datasets large enough to shuffle. Skew and spill do not appear on a thousand rows, and they are exam topics.

What to skip

Braindumps.

Pipeline and platform material. That belongs to the Data Engineer Associate exam. This one is about the API and the engine.

RDD-era tutorials. Plenty of older Spark material centres on RDDs. The exam is DataFrame-centric.

A sensible order

1. Read the exam guide and note the DataFrame API is 30%.
2. Work through the Spark programming guides.
3. Do the eight exercises above, reading the plan each time.
4. Take the practice test.
5. Return to the documentation for the weakest section.

Where this sits

The Databricks certification path explains why this exam is its own lane rather than a lighter version of data engineering — and why it is worth taking if the Spark fundamentals under the engineering exam felt shaky.