Skip to main content

Lesson 1 — What EasyFabric is

What you'll be able to do: explain, in one sentence, what EasyFabric produces and where the line sits between your job and its job.

The idea​

Imagine you want a house. One option is to lay every brick yourself. The other is to hand a builder a specification — three bedrooms, this layout, that kind of roof — and let them build it to a proven standard. You still make all the meaningful decisions; you just don't lay bricks.

EasyFabric is the builder. A complete data platform in Microsoft Fabric is normally weeks of manual, error-prone plumbing: creating lakehouses, defining tables in each layer, wiring up history tracking, generating SQL views, building a semantic model, setting up deployment pipelines. EasyFabric turns all of that into a specification you write once in YAML. Push the specification, and the platform builds itself.

This is the framework's first principle:

Framework over implementation. EasyFabric provides the tools and the standard; you provide the specifics — which sources, which tables, which columns, which business logic.

You are never editing the bricks (the generated Spark code, the table DDL, the semantic-model definition). You edit the specification, and the framework regenerates the bricks consistently every time.

In EasyFabric​

Concretely, your day-to-day work is just two kinds of files, written in VS Code:

  • YAML configuration — describes your sources, tables, columns, and semantic model.
  • Notebooks (.ipynb) — only for the genuinely custom bits (an unusual source, a special transformation). The common 99% needs no notebook at all.

Everything else — the lakehouses, the Bronze/Silver/Gold tables, the history tables, the SQL endpoint views, the semantic model, the Power BI wiring — is generated and deployed for you by Azure DevOps pipelines the moment you push.

So the honest one-sentence answer to "what is EasyFabric?":

You describe the data platform you want in YAML; EasyFabric generates and deploys a complete, standardized Bronze→Silver→Gold platform in Microsoft Fabric.

Check your understanding​

Answer each before expanding it.

A colleague says "EasyFabric is a tool that writes my Spark transformation code for me, so I should tune the generated code." What's off about that framing?

You don't tune the generated code — that's an implementation detail the framework owns and regenerates. You change the specification (the YAML), and let it regenerate. Editing generated output directly would be undone on the next deploy, and it breaks the "framework over implementation" contract. Custom logic goes in your notebooks or YAML expressions, which the framework calls at defined points — not in its output.

You need to add a brand-new data source that's unusual (say, a niche API). Which part of the "framework vs. you" line does that fall on?

Sources are exactly where your specifics live. The common source types are handled by the framework with just YAML. A genuinely unusual source is the one case where you write a small notebook connector — but you still describe the resulting table in YAML, so it gets the full medallion treatment. See Lesson 6 and the custom source topic.

Recap​

  • EasyFabric is a framework: you write a YAML specification, it builds and deploys the platform.
  • Framework over implementation — you own the what (sources, tables, logic); it owns the how (generated code and objects).
  • Your two tools are YAML and, occasionally, notebooks. You never hand-edit generated output.

Next: the shape of the platform it builds — The medallion architecture →