// the take
A grab-bag of Databricks-authored example notebooks for Delta Live Tables (DLT), the declarative pipeline framework now folded into 'Lakeflow Declarative Pipelines'. It's not a library — it's a collection of unrelated demos (Wikipedia clickstream, CDC via DMS+Terraform, Kafka streaming, loan-risk ML, Twitter sentiment) meant to be copied into a Databricks workspace and run as-is. Useful if you're evaluating DLT and want copy-paste starting points rather than reading the docs cold.
The CDC example (change-data-capture-example) actually shows the APPLY CHANGES INTO pattern end-to-end with both SQL and Python versions side by side, which is the one DLT feature that's genuinely hard to find good examples for elsewhere. The applyInPandasWithState notebook on stateful time-weighted averages is a solid, non-trivial streaming pattern most tutorials skip. The dms-dlt-cdc-demo folder includes full Terraform modules for standing up RDS MySQL + DMS + the pipeline, not just the notebook — rare for a vendor examples repo to go that far.
Every notebook requires a Databricks workspace and compute to do anything — there's no way to run or even lint this locally, and no CI. DLT itself was rebranded to Lakeflow Declarative Pipelines; some of these examples predate that and use syntax/APIs that may already be deprecated without any note of which is current. It's an accumulation of demos from different teams over years (Twitter sentiment via a dead API, hardcoded `/Repos/michael@databricks.com/...` paths in the README) rather than a maintained set — no versioning tells you which examples still reflect current DLT behavior and which are stale. No tests, obviously, and no indication anyone's fixing broken ones since last push was over a year ago relative to this kind of fast-moving platform feature.
We read the README, the directory tree and a key source file in November 2026. Nobody pays to be listed. How we judge →