Classification Drills & Keys
Classify each without looking at the key.
- Minimal set of attributes that uniquely identifies an entity instance.
__________ - Candidate key chosen as unique identifier.
__________ - Composite key whose components are foreign keys.
__________ - System-generated meaningless technical identifier.
__________ - How many instances participate in a relationship.
__________ - Number of entity types in a relationship.
__________ - Complete valid value set for an attribute.
__________ - Measured numeric rows in a dimensional structure.
__________ - Meaning of a single fact-table row.
__________ - Preserve changing dimension history by adding a new row.
__________ - High-level model of critical concepts and relationships.
__________ - Detailed requirements model that is technology-independent.
__________ - Technology-specific implementation model.
__________ - Message/payload structure used between systems.
__________ - Split a table by columns.
__________ - Split a table by rows.
__________ - Deliberately add redundancy for physical performance/security.
__________ - Remove repeating groups and ensure atomic attributes.
__________ - Move common features into a supertype.
__________ - Start from an existing database and document upward.
__________ - Resolve M:M with a relationship entity.
__________ - Reusable generic modeling structure.
__________ - Pre-built broad model for an industry.
__________ - Database-design acronym for Performance/ease, Reusability, Integrity, Security, Maintainability.
__________ - Quality category asking whether model and actual stored data agree.
__________
Classification Key
- Candidate key
- Primary key
- Compound key
- Surrogate key
- Cardinality
- Arity
- Domain
- Fact table
- Grain
- SCD Type 2
- Conceptual Data Model
- Logical Data Model
- Physical Data Model
- Canonical model
- Vertical partitioning
- Horizontal partitioning
- Denormalization
- First Normal Form (1NF)
- Generalization
- Reverse engineering
- Associative entity
- Data model pattern
- Industry data model
- PRISM
- Metadata matches the data
Changed-fact extension
After each correct classification, change one fact so a neighboring answer becomes correct. Examples:
- Candidate → Primary: state that the candidate was selected.
- Vertical → Horizontal partition: change “columns” to “records by year.”
- CDM → LDM: add attributes/domains/keys but keep it technology-independent.
- LDM → PDM: add chosen-DBMS datatypes/indexes/partitions.
- Standard view → Materialized: require pre-instantiated scheduled result.