finds.dev

jrlasak/databricks-code-practice

Practice Databricks coding skills with hands-on exercises. Import into Databricks Free Edition, write code, run assertions, check pass/fail. Covers Delta Lake, Spark SQL, PySpark, Auto Loader, medallion architecture, window functions, and more.

View on GitHubHomepage ↗

★ 294PythonCC-BY-SA-4.0updated Oct 2026

// the take

A structured set of hands-on Databricks exercises, pipeline labs, and deep-dives for data engineers who want to build muscle memory in Delta Lake, Spark SQL, PySpark, Auto Loader, and Structured Streaming rather than watch tutorials. Built for Databricks Free Edition, so it targets people studying for the DE Associate or PySpark Developer Associate certs, or anyone who wants assertion-checked reps instead of passive content.

148 exercises with assertion-based pass/fail checks force you to actually write working code instead of reading along; the three-tier structure (atomic exercises -> multi-notebook pipeline labs -> focused deep-dives) is a sensible progression from drilling a single concept to assembling a real medallion pipeline. The pipeline labs ship their own synthetic data generators, so the 500K+ transactions/day fintech lab or the streaming retail pipeline aren't toy datasets with five rows. Topic coverage is broad and current for the certs (liquid clustering, deletion vectors, CDF, watermarked dedup, stream-stream joins) and everything runs on the free tier with no cluster config or cloud account needed.

Solutions are checked directly into the repo next to the exercises, so there's no enforced discipline stopping you from peeking before you've actually struggled with the problem - it relies entirely on your own willpower. There's no CI running the exercises/assertions against a live workspace, so correctness depends on user-reported GitHub issues catching breakage, which is a slow feedback loop for 19 notebooks' worth of content. The README is heavily wrapped in lead-gen (UTM-tagged links to a newsletter, podcast, and paid interview-prep product), which doesn't affect the exercises themselves but signals the repo is a funnel for the author's other products as much as a standalone resource. Everything is tuned to Free Edition's serverless/Unity-Catalog defaults, so habits around cluster sizing, job scheduling, secrets, and networking that matter on a real paid workspace aren't covered at all.

We read the README, the directory tree and a key source file in November 2026. Nobody pays to be listed. How we judge →

// at a glance

Stars
294
Forks
166
Language
Python
License
CC-BY-SA-4.0
Last commit
19 Oct 2026

// nearby

If databricks-code-practice isn't quite it.

All repos →