Rapid Recall — R36–R70
Prompts
- Name four modeling deliverables.
- Forward engineering sequence?
- Reverse engineering sequence?
- Name five conceptual-model build steps.
- Name six logical-model build steps.
- Name eight physical-model build/adaptation steps.
- Name six modeling tools/resources.
- Recite PRISM.
- Name three model-governance control areas.
- What five change-history questions should version control record?
- Name the ten Data Model Scorecard categories.
- Model vs diagram?
- Unary hierarchy vs unary network?
- Dependent vs independent entity?
- Give five domain forms and what each constrains.
- Star vs snowflake?
- Conformed dimension vs conformed fact?
- Reconstruct Data Vault hub/link/satellite and Anchor-modeling components.
- What extra behavior/access ideas distinguish OO/UML class modeling?
- What is the defining representation style of Fact-Based Modeling?
- View vs materialized view?
- Vertical vs horizontal partitioning?
- What problems do BCNF, 4NF, 5NF address at recognition level?
- What makes an associative entity necessary/useful?
- Subtype absorption vs supertype partition?
- What business rule must be preserved when a surrogate replaces a natural/composite PK?
- Name four core modeling deliverables beyond “the picture.”
- What is a data model pattern?
- What is an industry data model, and what must happen before local use?
- Logical naming vs physical naming?
- Distinguish modeling, lineage, profiling, Metadata-repository, pattern, industry-model resources.
- Model review vs model maintenance?
- What does the Data Model Scorecard test beyond diagram readability?
- Why is source-target mapping/lineage a modeling deliverable?
- Name major physical-design adaptation decisions and explain why they are requirement-driven.
Reconstruction Key R36–R70
- Diagram/model, definitions, issues/outstanding questions, lineage.
- Requirements → CDM → LDM → PDM.
- Existing DB → PDM → LDM → CDM.
- Scheme; notation; initial user-view CDM; reconcile enterprise terminology/rules; sign-off.
- Requirements; existing artifacts; associative entities; atomic attributes; domains; keys/detailed relationships.
- Resolve abstractions; add physical attribute details; reference data; surrogate keys when needed; denormalize if justified; index; partition; views/materialized structures.
- Modeling tools; lineage tools; profiling tools; Metadata repositories; model patterns; industry models.
- Performance/ease of use; Reusability; Integrity; Security; Maintainability.
- Standards; design reviews; version/change control.
- Why; What/How; When; Who; Where.
- Requirements; completeness; scheme/level fit; structural soundness; generic structures; naming; readability; definitions; enterprise consistency; Metadata matches data.
- Model = full representation; diagram = one visual view.
- Hierarchy = same-type instance has at most one parent; network = may have multiple parents.
- Dependent identity relies on parent-contributed key attribute(s); independent can be uniquely identified without that identifying dependency.
- Type = basic kind; format = pattern; list = enumerated set; range = boundaries; rule-based = conditional validity.
- Star = collapsed dimensions; snowflake = normalized dimension component/hierarchy tables.
- Conformed dimension = shared descriptive context; conformed fact = shared standardized measure.
- Data Vault: hubs, links, satellites. Anchor: anchors, attributes, ties, knots.
- Attributes plus operations/methods, encapsulation and access visibility such as public/internal/private.
- Objects, facts and roles, often verbalized in controlled natural language; attributes not primary construct.
- Standard view derives on request; materialized view stores an instantiated result refreshed at a predetermined time.
- Vertical splits columns; horizontal splits rows by values/ranges.
- BCNF strengthens functional-dependency rules; 4NF addresses multivalued dependencies; 5NF addresses join dependencies.
- It resolves M:M or higher-arity relationship and can carry facts about the relationship itself.
- Absorption = one supertype table with subtype-specific nullable columns; partition = separate subtype tables with inherited attributes.
- Preserve original natural/business uniqueness as alternate key/constraint if still a business requirement.
- Model/diagram; definitions; issues/questions; lineage/source mapping.
- Reusable generic modeling structure for recurring situations.
- Broad pre-built industry reference; customize/validate against local requirements before use.
- Logical favors business-readable full terms; physical adapts to DBMS conventions/constraints with traceability.
- Modeling tools create/manage models; lineage traces source/target; profiling reveals actual data; Metadata repository preserves shared knowledge; patterns/industry models provide reusable starting content.
- Review = point-in-time correctness/completeness/consistency; maintenance = ongoing synchronization as requirements/implementation change.
- Requirements capture, completeness, scheme fit, structural soundness, generic structures, naming, readability, definitions, enterprise consistency, Metadata-data agreement.
- It identifies where modeled data originates/how it maps to implementation and supports validation + impact analysis.
- Resolve abstractions; physical details/reference data; key implementation; denormalization; indexes; partitions; views — selected from actual requirements/performance/security/access constraints, not habit.
← R1–R35 · Rebuild Drills →