MicroStrategy Developer key terms

A governed abstraction between physical warehouse tables and the objects report authors use, defining entities, metrics, and relationships once so every consumer gets consistent, reusable business logic instead of rewriting SQL.

The top-level container in a project that maps a set of logical tables and relationships to physical warehouse tables, including join specifications, attribute forms, and fact definitions.

A business entity used for grouping and filtering, such as Customer or Month; it is defined by one or more forms and resolves to lookup and ID columns in the warehouse.

A named column expression providing one piece of information about an attribute, for example Customer ID (the key) or Customer Name (a display description).

A single member value of an attribute, such as the customer 'Acme Corp'; elements are the values that appear as row headers and filter selections in reports.

An attribute whose identity is determined by a combination of two or more other attributes, used when no single natural key uniquely identifies the entity.

A declared dependency stating that one attribute determines another, such as City determining State; it drives automatic drilling and controls how attributes may be combined in one report.

A navigation hierarchy derived from attribute relationships that lets users move from a summary level to more detail, for example Region to State to City.

A numeric, additive measure stored at a declared grain in the warehouse, such as Revenue or Quantity; facts become the raw inputs that metrics aggregate or transform.

The set of attributes at which a fact is stored, telling the engine the lowest grain available so it never requests data at a level the warehouse cannot supply.

A governed analytical calculation built from facts, attributes, and functions, saved centrally so every report or API consumer uses the same formula and aggregation rules.

A metric whose aggregation is explicitly tied to a specific attribute level, such as Revenue at Customer, so results stay stable regardless of what other attributes appear on the report.

A metric defined by combining other metrics with arithmetic, functions, or conditionals, for example Profit Margin as Profit divided by Revenue.

The attribute grain at which a metric's value is calculated and returned, determined by its definition, dimensionality, and the attributes present in the report.

A saved or in-report restriction that limits which attribute elements or metric values are returned, evaluated in the warehouse or by the engine depending on type.

A parameter presented to the user or API caller at run time, supplying attribute elements, metric thresholds, or object choices that shape the query without editing the report.

A single certified version of a metric, attribute, or filter that is owned, reviewed, and reused across projects, preventing conflicting calculations from spreading into dashboards and downstream tools.

The controlled promotion of semantic objects between projects or environments using packages or update schemes, preserving dependency integrity and version traceability.

A multi-pass, multi-source in-memory dataset built from the semantic layer that pre-computes and stores results for fast dashboard and mobile consumption, refreshed on a schedule or on demand.

A dataset assembled and held in the analytics server's memory from one or more sources, letting the engine combine tables that do not share a common warehouse join.

A connection mode where queries are passed to the source database at run time rather than served from a stored dataset, trading speed for real-time freshness.

A stored connection definition holding driver, server, port, authentication, and connection-pool settings that semantic objects use to reach a specific warehouse.

The binding of a project's logical tables to a database instance and physical namespace, allowing the same semantic objects to run against different environments.

A control that inspects generated SQL before execution and can block, warn about, or limit queries by cost, row count, or banned patterns to protect warehouse resources.

The persisted result set of a published cube, served to users without re-querying the warehouse until a refresh, schema change, or security change invalidates it.

The deliberate or automatic discarding of cached results when underlying data, object definitions, or security assignments change, forcing regeneration on the next request.

A mandatory filter attached to a user or group that restricts which warehouse rows their queries can return, applied beneath any report-level filtering.

An access pattern where different users see different rows of the same report based on their identity, typically implemented through security filters or database-level policies.

A granular permission granting a user or group the ability to perform an action, such as creating a metric, publishing a cube, or using a specific editor.

The per-object list of users and groups with granted or denied permissions, determining who may view, modify, or delete a semantic object or folder.

A named bundle of privileges and object permissions assigned to users so that access is managed consistently rather than user by user.

A recorded trail of system and user activity such as logins, object changes, and job executions, used for troubleshooting, compliance evidence, and forensic review.

An open protocol that lets AI assistants discover and invoke external tools and data sources through standardized server endpoints, exposing governed metadata and query capabilities to a model.

A service that publishes resources, prompts, and callable tools over the Model Context Protocol so an AI client can retrieve governed definitions and execute approved analytics operations.

A structured request from an AI model to an MCP server naming a tool and its arguments, whose result is returned to the model as context for the next reasoning step.

An HTTP interface using resource-oriented URLs and standard verbs that lets external applications authenticate, browse the metadata, and execute reports or cubes programmatically.

A time-limited credential issued after login that authorizes subsequent API calls, carrying the caller's identity so server-side security filters and privileges still apply.