Skip to main content

Lesson 10 — Where to go next

What you'll be able to do: hold the whole EasyFabric mental model in one picture, prove it to yourself, and pick your next reading by role.

The whole picture in one breath​

You declare a platform in YAML — connections, objects, a model. On push, an Azure DevOps pipeline validates it and the generator deploys a full Bronze → Silver → Gold medallion in Fabric, with every object created empty (schema-on-write). Then, at run time, the wheel package — driven by the same YAML — loads data: raw into Bronze (with history), cleaned into Silver, and shaped into a Gold star schema that becomes a semantic model for Power BI.

If that paragraph reads as obvious rather than dense, the course did its job. Every noun in it was a lesson.

Final self-test​

No hints this time — say each answer out loud before checking. Being able to generate these from memory is the difference between "I read a course" and "I understand this."

1. Why does deploy create empty tables instead of tables with data?

Because EasyFabric is schema-on-write: structure is decided and built up front, and loading is a separate step. Deploy makes the shape exist; the run-time loader fills it. (Lessons 3, 9)

2. Where does the same YAML get read twice, and by what?

By the generator at deploy time (to build object shapes) and by the wheel package at run time (to load data). One source of truth, two consumers — the heart of "metadata-driven." (Lesson 4)

3. A row appears wrong in a report. Which layers can you walk back through, and what does each hold?

Gold (business-shaped) → Silver (clean, typed, quality-checked) → Bronze (raw as delivered, plus history). Traceability is a core reason the layers exist. (Lesson 2)

4. You add a new table. What has to change in the loader and the pipeline?

Nothing in either — both are metadata-driven. Declare the object in YAML; the loader picks it up and the pipeline deploys it automatically. (Lessons 3, 7, 9)

5. What's the difference between a fact and a dimension, and where do both come from?

A fact is a measurement you sum; a dimension is context you slice by. Both are Silver tables, arranged into a star by the Gold model. (Lesson 8)

If any answer didn't come, that lesson is your next re-read — that's the self-test doing its job.

Choose your path​

You now have the mental model. Where you go deeper depends on your role:

Building the platform (data engineer)

Modeling for analytics (analyst / BI developer)

Setting up and shipping (platform / DevOps)

Working with Claude Code

  • Skills — add-source, tabular-developer, notebook-developer and more do the mechanical work while you stay in the concepts you just learned.

The one thing to remember​

Every question — "where does this go?", "why is this empty?", "how do I change that?" — resolves the same way: it's all driven by the YAML. Describe what you want; the framework builds it. That's EasyFabric.

← Back to the course overview · or jump into the Basics tutorials.