Fundamentals of Microsoft Fabric (for True Epub) (Nikola Ilic, Ben Weissman)（Z-Library）

# Fundamentals of Microsoft Fabric ## 【One-Line Pitch】 A practical, hands-on guide to Microsoft's unified data platform, showing how Fabric's integrated workloads—from lakehouses and warehouses to real-time analytics and Power BI—can replace fragmented data stacks. Essential reading for data engineers, analysts, and Power BI professionals who want to navigate Fabric's many choices without getting lost. ## 【Book Arc】 - **Opening (~0%–4%)**: Introduces Fabric as Microsoft's "biggest data product after SQL Server," positioning it as a unified SaaS platform that consolidates multiple data workloads. Covers tenant setup, capacity management, and the foundational concept that Fabric intentionally overlaps features across roles to bridge knowledge gaps. - **Early (~4%–14%)**: Walks through creating a lakehouse, understanding OneLake's metadata layers (technical, business, operational), and connecting external sources like Snowflake, Amazon S3, and on-premises systems. Establishes the lakehouse as the centerpiece of Fabric's architecture. - **Early (~14%–25%)**: Explores Data Factory pipelines and Dataflow Gen 2 for no-code/low-code data movement and transformation. Introduces the medallion architecture (bronze/silver/gold) using the New York Taxi dataset as a running example, then transitions into data warehousing fundamentals. - **Early–Middle (~25%–39%)**: Covers Fabric warehouse capabilities—zero-copy table cloning, time travel, and schema-on-write—alongside data science features like SynapseML and AutoML. Moves into Real-Time Intelligence with KQL databases, streaming connectors (Kafka, CDC), and Real-Time Dashboards. - **Middle (~39%–50%)**: Deep-dives into Power BI integration, focusing on Direct Lake mode—the feature that delivers Import-mode performance without copying data. Explains transcoding (loading Delta table columns into cache), limitations (no DAX calculated columns, unenforced primary keys), and how Fabric changes the semantic model landscape. - **Late (~50%+)**: Covers SQL databases in Fabric with DevOps workflows, CI/CD integration, autoscaling, and the built-in GraphQL interface for flexible querying. Ties together how operational and analytical workloads can coexist in one platform. ## 【Key Takeaways】 - **Fabric is a unified SaaS platform, not another tool to bolt on** (Opening): Microsoft deliberately overlaps features across data roles to bridge knowledge gaps and increase transparency across the tech stack. This means learning one workload often transfers to others. - **The lakehouse is the architectural foundation** (Early): It stores raw data in native formats (JSON, images, videos, logs) while supporting schema enforcement, ACID transactions, versioning, and advanced analytics—all in OneLake. This eliminates the traditional split between data lakes and warehouses. - **Medallion architecture is the recommended design pattern** (Early): Organizing data into bronze (raw), silver (cleaned), and gold (curated) layers within lakehouse schemas provides a clear, scalable path from ingestion to analytics-ready data. - **Pipelines and Dataflow Gen 2 cover the movement spectrum** (Early): Pipelines handle complex orchestration with activities, triggers, and control flows; Dataflow Gen 2 offers no-code visual transformations. Both integrate tightly with Fabric's ecosystem, reducing configuration overhead. - **Fabric warehouses add modern capabilities beyond traditional SQL** (Early): Zero-copy table cloning enables instant, storage-efficient development and testing environments, while time travel lets you query data as it existed at any timestamp—both via simple T-SQL. - **Direct Lake is the killer feature for Power BI professionals** (Middle): It delivers Import-mode performance without copying data by transcoding only needed Delta table columns into cache memory. However, it requires moving calculated column logic upstream and accepting that constraints aren't enforced. - **Real-Time Intelligence is now a first-class workload** (Middle): KQL databases (SaaS evolution of Azure Data Explorer) handle telemetry, logs, and time-series data with optimized storage and indexing. Real-Time Dashboards and Activator alerts enable operational monitoring without leaving Fabric. - **SQL databases in Fabric embrace modern DevOps** (Late): Built-in CI/CD, source control, autoscaling, and a GraphQL interface make Fabric SQL suitable for both operational and analytical workloads, with real-time replication to Power BI. ## 【Reading Tips】 - **Skim the opening chapters** (~0%–4%) if you already understand Fabric's value proposition; focus instead on the capacity settings and tenant configuration details, which are practical prerequisites. - **Deep-read the lakehouse and medallion architecture sections** (~7%–25%): The New York Taxi dataset walkthrough is the book's core hands-on example—follow along in your own Fabric tenant to internalize the workflow. - **Pay special attention to the Direct Lake chapter** (~46%–50%): This is where the book earns its keep for Power BI professionals. The limitations (no DAX calculated columns, unenforced primary keys) are critical gotchas that will save you hours of debugging. - **Treat the Real-Time Intelligence chapters** (~36%–39%) as an overview rather than a deep reference: KQL query syntax and dashboard configuration are covered, but you'll likely need Microsoft's docs for production-scale implementations. - **Use the summary sections as review checkpoints**: Each chapter ends with a summary that distills the key decisions and trade-offs—read these first if you're short on time, then go back for details. ## 【Coverage Limits】 This guide synthesizes the first half of the book (through ~50%), covering lakehouse fundamentals, data movement, warehousing, real-time analytics, and Power BI integration. Later chapters on SQL databases, GraphQL, and advanced scenarios are only partially covered in the available excerpts. ##