Building Medallion Architectures (Piethein Strengholt) (z-library.sk, 1lib.sk, z-lib.sk)

【One-Line Pitch】 A field-tested guide to designing and running a medallion (bronze/silver/gold) lakehouse architecture on Azure Databricks and Microsoft Fabric, written for data engineers, architects, and data leaders who need to move past the hype and build pipelines that actually support decisions. 【Book Arc】 - **Opening (~0%–10%)**: Sets the stakes — medallion architectures are widely used but widely misunderstood, often mistaken for a rigid step-by-step recipe rather than a flexible design standard. Traces the big-data-to-lakehouse history (data lakes, the "data swamp" backlash, relational discipline) to explain why the pattern exists. - **Early (~10%–35%)**: Frames the book's purpose and audience: a practical reference spanning novices to seasoned professionals, with a survey-driven emphasis on governance, security, data mesh, and generative AI. Outlines the four-part structure and the vendor-neutral intent behind the Azure/Fabric examples. - **Middle (~35%–55%)**: Part I builds theory — the evolution of data architecture, enduring design principles, and a three-layered model (providers → distribution platform → consumers, wrapped in metadata and governance), then breaks down each medallion layer's function and interactions. - **Late (~55%–85%)**: Part II turns hands-on, walking through building the medallion layers end to end in Microsoft Fabric, with an external Databricks path for readers on that stack; concepts are presented vendor-neutrally so skills transfer to other Spark/Delta Lake platforms. - **Ending (~85%–100%)**: Closes on scaling and modern trends — data mesh, governance, data contracts, security, and AI — plus real-organization case material (e.g., AP Pension, Amadeus) on what worked and what didn't. (Excerpts do not cover the closing chapters in detail.) 【Book Arc note】The excerpt set is front-loaded (preface, TOC, early chapters); later implementation chapters are only summarized, not sampled. 【Key Takeaways】 - **Medallion is a flexible design standard, not a rigid framework** (Opening): treating bronze/silver/gold as a fixed recipe causes misinterpretation and inefficiency; the book's core move is explaining the "why" behind each pattern so you can adapt it. - **History explains today's architecture** (Middle): the three-layered provider/distribution/consumer model with an overarching metadata-and-governance layer recurs throughout; understanding warehouse discipline versus lake innovation clarifies current trade-offs. - **Context beats dogma** (Early): the foreword stresses that documented "right ways" lose value without the reasoning; readers should decide which processes fit their organization and skip the rest. - **Vendor-neutral skills, concrete tools** (Early): Fabric and Databricks are teaching vehicles, not promotions; Airflow, Data Factory, PySpark, and metadata-driven methods are chosen because they transfer across Spark/Delta Lake platforms. - **Governance and security are first-class** (Early): practitioner surveys flagged these as top expectations, and they thread through the architecture rather than being bolted on. - **Modern trends are part of the design conversation** (Late): data mesh, data contracts, and AI are treated as scaling questions for medallion architectures — e.g., whether mesh replaces a lakehouse — not separate topics. - **Real organizations inform the guidance** (Ending): case input from companies like AP Pension and Amadeus grounds claims about what has and hasn't worked in practice. 【Reading Tips】 - **Skim Part I if you come from data warehousing**: the historical and conceptual chapters will feel familiar; slow down on Chapter 3's layer breakdown, which anchors everything later. - **Deep-read Part II with your own stack open**: follow the Fabric walkthrough hands-on, and pull the companion Databricks blog post if that's your platform. - **Treat it as a reference, not a cover-to-cover read**: the foreword explicitly frames it as a companion to revisit during implementation. - **Watch for the "why" passages**: when the author explains a decision's rationale, note it — that's the transferable part when your context differs. - **Jump to governance/security and trends chapters early if you're a CDO or architect**: they answer the questions most often raised in design workshops. 【Coverage Limits】 This guide is based on stratified excerpts weighted toward the preface, table of contents, and opening chapters; the hands-on implementation chapters and closing case studies are only summarized, so specific techniques, code, and chapter-level detail there are not covered.