MicroStrategy Analyst flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Semantic layer

A metadata model that maps warehouse tables to business attributes, facts, and metrics so the platform generates SQL instead of requiring hand-written queries.

Attribute ID form

The key form of an attribute used for joins and GROUP BY. Reports always aggregate on the ID form, while the DESC form supplies display labels.

Attribute relationship

A declared 1-to-many lookup between child and parent attributes, such as City to State. It enables automatic drilling and correct aggregation rollups.

Fact

A numeric value stored at a defined attribute grain, such as Revenue at Day and Item. Facts are the raw inputs from which metrics are built.

Fact extension

A calculation added to a fact without altering the warehouse table: conditional, logical, or derived. Implicit facts are created automatically from table columns.

Metric dimensionality

The set of attributes at which a metric is calculated. It determines which relationship joins the engine adds to allocate values to report rows.

Level metric

A metric evaluated at a specified target attribute level before being combined with report rows, used to avoid incorrect re-aggregation of ratios.

Compound metric

A metric built from arithmetic on other metrics. It inherits the dimensionality of its inputs and is stored in the project schema.

Derived metric

A metric defined at the report or dossier level rather than in the schema. It is not reusable across reports and is not governed centrally.

Governed metric

A single certified definition with formula, filter, dimensionality, and format stored once in the schema and reused by every report and dataset.

Metric filter

A condition embedded in the metric definition that restricts which fact rows are aggregated, distinct from a report filter that limits returned rows.

Non-additive metric

A metric such as a distinct count or ratio that cannot be summed across rows. It must be defined at the correct level to produce valid totals.

Smart metric

A metric the warehouse computes at its native level, with the engine re-aggregating results. Valid only when the aggregation is decomposable.

Metric certification

The workflow of validating a metric against source totals, approving it by a data steward, and publishing it to shared folders for governed reuse.

Definition drift

Divergence that occurs when format, filter, or formula overrides are applied outside the metric object, causing different consumers to see different numbers.

Metric lineage

The traceable path from metric to fact or column, to table, to source system, including stated grain, filters, and numerator and denominator logic.

Database instance

A connection object storing driver, server, port, credentials, and connection-mapping settings. Multiple instances can target the same warehouse for different workloads.

Connection mapping

An assignment of different database instances to users or groups, evaluated at query time, so analysts can hit a read-only replica while others use the primary.

Intelligent Cube

A pre-computed multi-dimensional data set stored in the Intelligence Server. Queries are answered from the cube rather than the warehouse, trading freshness for speed.

Incremental cube refresh

A refresh mode that appends only new data to an existing cube, as opposed to a full rebuild, reducing load and refresh duration.

Report cache

A stored prior result set keyed by generated SQL and security profile. It is invalidated on schema change, cube refresh, or manual cache clear.

In-memory dataset

Imported data cached on the server for a dossier, refreshed on a schedule, with row limits controlling staleness and memory footprint.

VLDB settings

Warehouse-specific query options such as row limits, timeouts, and intermediate table creation, configured at the database instance level to prevent runaway SQL.

Free SQL

Direct hand-written SQL access that bypasses the semantic layer. It is fast to build but loses governed metric definitions and row-level security enforcement.

Security filter

A schema-level object assigned through user or group ACLs that appends a row-level WHERE predicate to every query for that user.

Project-level ACL

Permissions such as Browse, Use, Create, and Configure granted at the project level, forming the first layer of the access-control model.

Object-level ACL

Permissions applied to individual folders, reports, and dossiers, controlling who can view, modify, or delete each object.

Attribute-based access

A data-security technique that denies specific elements of an attribute, such as hiding certain cost centers, rather than filtering whole rows.

AI security inheritance

The rule that AI and natural-language query features run under the requesting user's security profile, never a broader service account.

AI exposure governance

Restricting which datasets and certified metrics are available to AI assistants, and logging prompts and responses for audit review.

Least privilege

Granting permissions through group roles rather than per-user object ACLs, auditing inherited rights, and removing default Everyone grants on sensitive folders.

Audit log review

Using enterprise logs of logins, object access, and query history, with alerts on failed logins and unusual exports, to support access-control review.