7.1. Terms A–Z
Analyze feature: Right-click analysis that explains increases or decreases and finds where distributions differ (4.3.1). · Anomaly detection: Analytics pane option that flags unexpected points on time-series line charts with possible explanations (4.3.3). · App: Packaged, read-only distribution of workspace content with audiences and controlled updates (5.1.2). · Append Queries: Power Query operation that stacks tables vertically, matching columns by name (2.3.5). · Automatic page refresh: Page-level refresh of visuals for DirectQuery, Direct Lake, and live sources, by fixed interval or change detection (4.2.5). · Auto date/time: Option that creates hidden date tables per date column; usually disabled in favor of a common date table (3.1.4).
Bidirectional filtering: Cross-filter direction Both, letting filters flow both ways across a relationship (3.1.2). · Binning: Grouping continuous values into equal-size ranges (4.3.1). · Bookmark: Saved page state (data, display, page, selected visuals) (4.2.1). · Build permission: Semantic model permission to create reports, composite models, and Analyze in Excel connections (5.2.2).
CALCULATE: DAX function that modifies filter context before evaluating an expression; boolean filters overwrite existing column filters (3.2.2). · Calculated column: DAX column computed per row at refresh and stored (1.3.2, 3.1.5). · Calculated table: DAX table computed at refresh and stored (3.1.5). · Calculation group: Set of calculation items applying reusable logic to any explicit measure via SELECTEDMEASURE (3.2.5). · Cardinality (relationship): One-to-many, one-to-one, or many-to-many (3.1.2). · Cardinality (column): Number of distinct values; drives model size (1.2.2, 3.3.1). · Certified: Endorsement applied by admin-authorized reviewers (5.1.3). · Clustering: Automatic segmentation of items on a scatter chart (4.3.1). · Column distribution / profile / quality: Power Query profiling views (2.2.1). · Composite model: Model combining storage modes or sources, including DirectQuery to a Power BI semantic model plus local tables (2.1.1). · Conditional formatting: Value-driven formatting by gradient, rules, or field value (4.1.2). · Context transition: Conversion of row context into filter context by CALCULATE or a measure reference (1.4.1, 3.2.2). · Copilot: AI assistant in Power BI requiring F2/P1+ capacity or a Fabric Copilot capacity (4.1.4, 4.3.2).
DAX: Data Analysis Expressions, the model's formula language (1.3.2). · DAX query view: Desktop view for writing and running DAX queries (3.3.2). · Decomposition tree: AI visual for drilling a measure across dimensions with AI splits (4.3.2). · Dimension table: Descriptive table on the one side of relationships (1.2.1). · Direct Lake: Fabric storage mode reading OneLake Delta tables into memory; framing instead of copy refresh (2.1.3). · DirectQuery: Storage mode that queries the source for every visual (2.1.3). · Drillthrough: Navigation to a target page filtered to the selected data point (4.2.3). · Dual: Table storage mode acting as Import or DirectQuery as needed (2.1.3). · Duplicate query: Independent copy of a query's steps (2.3.5).
Enable load: Query setting that controls whether output loads into the model (2.3.6). · Endorsement: Promoted, certified, or master data badges (5.1.3). · Error bars: Analytics pane bounds showing uncertainty (4.3.3).
Fact table: Table of events or measurements at a declared grain (1.2.1). · Field parameter: Report feature that lets users switch the measure or dimension in a visual (2.1.4, 4.1.1). · Filter context: The set of filters active when a measure is evaluated (1.4.1). · Fixed Decimal Number: Currency-friendly data type with four decimal places (2.3.1). · Forecast: Analytics pane projection on continuous-date line charts (4.3.3). · Framing: Direct Lake's metadata-only refresh (2.1.3).
Gateway (on-premises data gateway): Relay that lets the service reach on-premises or private sources; standard or personal mode (5.1.4). · Grain: What one row of a fact table represents (1.2.1).
Import: Storage mode that loads a compressed copy into VertiPaq (2.1.3). · Incremental refresh: Partitioned refresh of recent data using RangeStart and RangeEnd (2.1.4, 5.1.4).
KEEPFILTERS: CALCULATE modifier that intersects rather than overwrites filters (3.2.2). · Key influencers: AI visual ranking factors associated with an outcome (4.3.2).
LASTNONBLANK: Function returning the last date with data, used for semi-additive measures (3.2.4). · Live connection: Report connection to a shared semantic model without local modeling (1.1.2, 2.1.1).
Mark as date table: Setting that designates a contiguous, unique date column for time intelligence (3.1.4). · Measure: DAX calculation evaluated at query time in filter context (1.3.2, 3.2.1). · Merge Queries: Power Query join adding columns from a matching table (2.3.5). · Mobile layout: Phone-optimized arrangement of a report page (4.2.4).
Paginated report: Pixel-perfect, multi-page report authored in Power BI Report Builder (4.1.6). · Performance Analyzer: Tool that times DAX query, visual display, and other per visual (3.3.2). · Personalize visuals: Reader-side changes to visuals saved as personal bookmarks (4.2.4). · Power Query (M): Data preparation engine and language (1.3.1). · Privacy levels: Private, Organizational, Public settings that govern data combination across sources (2.1.2). · Promoted: Endorsement any writer can apply (5.1.3). · Publish to web: Public, unauthenticated embedding; never for internal data (5.1.2).
Q&A visual: Legacy natural-language visual retiring in February 2027 (4.3.2). · Query folding: Translation of Power Query steps into a native source query (1.3.1). · Quick measures: Dialog-generated DAX for common calculations (3.2.1).
RangeStart / RangeEnd: Date/Time parameters required for incremental refresh (2.1.4). · Reference query: Query whose source is another query's output (2.3.5). · Report page tooltip: A tooltip-sized page shown on hover (4.2.2). · Role-playing dimension: One dimension used in several roles for a fact (3.1.3). · Row context: Iteration over individual rows in calculated columns and iterators (1.4.1). · Row-level security (RLS): Model roles filtering rows per user; applies to Viewers and read-only consumers (5.2.3).
SAMEPERIODLASTYEAR: Time intelligence function shifting the current dates back one year (3.2.3). · Scheduled refresh: Service-configured data refresh, up to 8/day shared and 48/day Premium/Fabric (5.1.4). · Semantic model: Governed model of tables, relationships, measures, and security (1.1.2). · Semi-additive measure: Measure that aggregates across some dimensions but not time (3.2.4). · Sensitivity label: Purview classification with protection that persists on export (5.2.4). · Sort by column: Property sorting one column by another's values (3.1.1). · Star schema: Fact table surrounded by dimensions in one-to-many relationships (1.2.2). · Subscription: Scheduled email snapshot of a page, dashboard, or paginated report (5.1.3). · Sync slicers: Pane that syncs slicer selections and visibility across pages (4.2.3).
Theme (JSON): File defining colors and formatting defaults for reports (4.1.2). · Time intelligence: DAX functions computing over shifted or accumulated date ranges (3.2.3).
Unpivot Other Columns: Unpivot all columns except those selected, robust to new columns (2.3.2). · USERELATIONSHIP: CALCULATE modifier activating an inactive relationship (3.1.3). · USERPRINCIPALNAME: DAX function returning the signed-in user's UPN for dynamic RLS (5.2.3).
VertiPaq: Power BI's in-memory columnar engine (1.2.2). · Visual calculation: DAX calculation stored on a single visual and evaluated over its visual matrix (4.1.5).
Workspace: Collaborative container in the service with roles and a license mode (5.1.1). · Workspace roles: Admin, Member, Contributor, Viewer (5.2.1).