Overview
ISO/IEC TR 19583-21:2022 - "Information technology - Concepts and usage of metadata - Part 21: 11179-3 Data model in SQL" is a technical report that provides a concrete, SQL-based instantiation of the registry metamodel defined in ISO/IEC 11179-3. The document demonstrates how UML Class Diagram constructs from the 11179-3 conceptual metamodel can be translated into SQL (per ISO/IEC 9075) to produce a working metadata registry schema, including example SQL and implementation guidance.
Key topics and technical requirements
- Scope and purpose
- Presents a specimen instantiation of the ISO/IEC 11179-3 metamodel using SQL to promote consistent registry implementations.
- Mapping UML → SQL
- Conventions for naming (keywords vs. object names), mapping each UML class and composite datatype to SQL tables, and attributes to columns.
- Use of surrogate identifiers as primary keys since UML lacks an equivalent.
- Datatype translation
- Guidance for translating primitive datatypes (Boolean, Date, Datetime/TIMESTAMP, Integer, String/Text) and handling larger or complex content (VARCHAR, BINARY/CHARACTER LARGE OBJECT).
- Reference tables in the report (Table 1 & Table 2) show recommended SQL datatype mappings.
- Multiplicity and collections
- Options for multivalued attributes: SQL collection types (ARRAY/MULTISET) or separate characteristic tables with foreign keys.
- Associations and association classes
- Named associations are instantiated as dedicated tables; association classes also become tables.
- Inheritance / specialization strategies
- Two common SQL patterns: single table with a discriminator column, or separate superclass/subclass tables with one-to-one PK = FK links; cross-table constraints discussed.
- Constraints and obligations
- Guidance on UNIQUE, CHECK and cross-table constraints.
- Important note: attribute/association obligations (Mandatory/Conditional/Optional) are enforced only when an item’s Registration Status is Recorded or higher. The report recommends leaving database constraints permissive and enforcing some obligations at the registry application level.
- Practical examples
- Includes example SQL to instantiate the complete metamodel (Annex A).
Applications and users
- Implementers of metadata registries and catalogues
- Data architects, data modelers and DBAs designing metadata stores
- Software developers building registry applications that conform to ISO/IEC 11179
- Enterprise architects and standards bodies seeking a reference SQL instantiation
Practical benefits include a standardized blueprint for storing metadata, clearer interoperability between registries, and reduced ambiguity when turning UML conceptual models into relational schemas.
Related standards
Keywords: ISO/IEC TR 19583-21:2022, ISO/IEC 11179-3, SQL instantiation, metadata registry, UML to SQL mapping, metadata management, data model in SQL.
