Rapid Battle Cards 31–38
31 — Associative Entity vs Ordinary Entity
- Associative entity: resolves M:M (or higher-arity) relationship and can store relationship-specific facts.
- Clue: the relationship itself needs identity/data, e.g. Registration Date for Student–Course.
32 — Subtype Absorption vs Supertype Partition
- Absorption: one supertype table with subtype-specific nullable columns.
- Partition: separate subtype tables carry inherited supertype attributes.
- Clue: one broad table vs separate subtype tables.
33 — Natural Key Retained vs Surrogate PK
If a surrogate replaces a natural/composite PK, preserve original business uniqueness as an alternate key/constraint.
- Clue: technical identity must not erase a business uniqueness rule.
34 — Data Model Pattern vs Industry Model
- Pattern: reusable generic structure for recurring situations.
- Industry model: broad pre-built reference for an industry, requiring customization.
- Clue: generic structural solution vs industry-wide starting content.
35 — Logical Naming vs Physical Naming
- Logical: meaningful, business-friendly full terms.
- Physical: approved abbreviations/underscores/constraints of the DBMS.
- Clue: business readability vs implementation constraints.
36 — Model Review vs Model Maintenance
- Review: correctness/completeness/consistency quality checkpoint.
- Maintenance: keep governed model synchronized as requirements/implementation change.
- Clue: quality gate vs ongoing currency.
37 — Standards vs Design Review vs Version Control
- Standards: rulebook / expected modeling practice.
- Review: checkpoint on a specific model.
- Version control: history of what changed and why.
- Clue: rulebook vs gate vs history.
38 — Scorecard Completeness vs Structural Soundness
- Completeness: are required things present and inappropriate extras absent?
- Structural soundness: is the design structurally valid/buildable?
- Clue: missing/extra vs structurally invalid.
Rapid Discrimination Drill
Customer_IDis generated and meaningless; unique Email remains business-recognizable. Likely: surrogate primary key + business/alternate Email key.- “One row = one student registration per class per day.” Concept: grain.
- Detailed attributes/domains/keys but no DBMS specifics. Level: LDM.
- Existing database is being documented upward into business meaning. Direction: reverse engineering.
- Reviewer asks whether a relational/dimensional/NoSQL model follows the rules of its declared scheme. Scorecard: match the model to its scheme/level.