chapter two

2 Semantic model concepts

 

This chapter covers

  • Modeling relationships, table expansion, and data lineage
  • Exploring user-driven filters and real-time report interactivity
  • Navigating expanded tables and disconnected entities
  • Laying the foundation for model-aware, filter-driven DAX logic

From the visual flow and mental model we built in the previous chapter, it is clear that everything in Power BI operates through its semantic model. This isn’t just a background detail—it’s the core infrastructure that governs how Power BI interprets data, applies filters, and calculates results. If you come from an Excel background, this concept may take time to get accustomed to. In Excel, formulas evaluate over values shown on the visual. However, in Power BI, DAX is technically part of the data model. Results shown on the report visual are the final outcome of queries and expressions evaluated over the underlying data model. Whether you're dragging fields onto a visual, building relationships, or writing measures, your actions rely on the semantic model working silently and intelligently behind the scenes.

2.1 What is a semantic model?

2.2 From tables to unified models

2.3 Relationships in Power BI and DAX

2.4 Expanded tables: How DAX sees the model

2.4.1 Navigating an expanded table

2.4.2 Using expanded tables in DAX

2.4.3 The importance of expanded tables

2.5 Data lineage: Tracking a column’s origins

2.6 What can break lineage?

2.7 Semantic models as a unified interactive system

2.7.1 Complex semantic models: beyond the basics

2.7.2 Disconnected entities in a unified semantic model

2.7.3 Why disconnected tables work

2.8 Connected vs. disconnected recordsets in interactive mode

2.8.1 When lineage is lost

2.8.2 Reconnecting virtual tables with TREATAS()

2.9 Summary