Learn metadata-driven generation with a small AdventureWorks customer, address and order model. Open it in the DataM8 app or use the CLI to inspect and edit metadata, then generate model documentation, Databricks notebooks or a Power BI semantic model. The first exercise runs locally without a database or cloud account.
| I want to… | Begin here | First result |
|---|---|---|
| Learn visual modeling | Open the sample in the app | Change a Customer description and confirm it survives reload |
| Validate and generate from the command line | Local CLI setup | Generate and inspect a fresh output/docs/index.md |
| Build and deploy my own solution | Extend the sample, then deploy | Extend your model, validate and generate it, then deploy to your own environment |
| Change templates or target conventions | Template development and configured targets | Change one template and compare the generated output |
| Understand the example | Model documentation target and property references | Follow Customer through Stage, Core, Curated and Consumer |
Recommended first result: generate docs and open its index. Use a disposable
copy: autosave writes metadata, and generation can replace a target's output.
Install the stable app or CLI
for product use; the local project below pins a public CLI revision for reproducibility.
Clone or download the repository, copy the complete solution to an exercise directory, and run these commands there with uv and a compatible Python:
uv sync --frozen
uv run datam8 validate --solution ./ORAYLISDatabricksSample.dm8s
uv run datam8 generate docs --solution ./ORAYLISDatabricksSample.dm8sOpen output/docs/index.md; it should contain Customer and links to entity pages
and a relationship diagram. The SQL Server metadata connector is bundled with
the pinned CLI and installed through its sql extra. No connector checkout,
download token or source connection is needed for this exercise.
| Target | Source | Output |
|---|---|---|
| databricks | generate/databricks-lake |
output/databricks |
| powerbi | generate/powerbi-tabular |
output/powerbi |
| docs | generate/docs |
output/docs |
The model follows AdventureWorks customers, addresses and orders through Stage, Core and Curated, including DimCustomer, DimDate and FactSalesOrder. Consumer entities retain Customer, Date, Sales Order and Total Sales. The SQL example uses the standard SalesLT.Customer table; no custom language-specific views are required. Data endpoints and secret references are neutral placeholders.
A neutral Databricks bundle under deployment/ and the own-deployment guides
cover the next steps. Supply your own identity, compute, catalog and source
connection. Credentials and live environment bindings are not included. Local
generation does not establish successful deployment, refresh or processing.
- Sample documentation
- Central illustrated handbook
- Local checks and review requirements
- Source and credential policy
Follow repository instructions and run the relevant local checks during implementation. Propose changes through a feature PR; the maintainer merges manually. Report product issues in DataM8, with no credentials or live environment details.