Konceptualiųjų apribojimų transformacija į SQL kodą

In this paper, the method is proposed for transforming UML class diagrams with OCL constraints to relational database schemas, having advantages over “UML Profile for Databases” models. The proposed method consists of two phases supplementing each other: metamodel based transformations and pattern-b...

Full description

Bibliographic Details
Main Author: Armonas, Andrius
Other Authors: Maciulevičius, Stasys
Format: Dissertation
Language:Lithuanian
Published: Lithuanian Academic Libraries Network (LABT) 2005
Subjects:
MDA
MOF
SQL
OCL
UML
CWM
Online Access:http://vddb.library.lt/fedora/get/LT-eLABa-0001:E.02~2005~D_20050524_190132-41686/DS.005.0.01.ETD
Description
Summary:In this paper, the method is proposed for transforming UML class diagrams with OCL constraints to relational database schemas, having advantages over “UML Profile for Databases” models. The proposed method consists of two phases supplementing each other: metamodel based transformations and pattern-based transformations. All transformations are based on OMG official standards or RFPs (Request for Proposals) and are prepared for use in MDA (Model Driven Architecture) context. This means, that resulting models, created using the described method, are long-lasting, independent from platform and abstract enough to be independent from technology. This work covers analysis of the field of object – relational transformations, analysis of support of OCL in currently used UML tools, transformation rule sets of metamodel transformations and pattern-based transformations, a lot of examples illustrating every transformation. Metamodel transformations are designed to transform object models written in UML language to corresponding relational models: types, classes, attributes, association classes, associations to tables, columns, foreign keys and other concepts. Pattern-based transformations are designed for generating static relational concepts such as check constraints, views, stored procedures and triggers. Using the two described transformation types and the whole method described, one can generate full-fledged relational database schemas, have precise UML models and keep much more... [to full text]