InformatieAnalyse 0 : opereren op conceptueel niveau

Transcription

1 InformatieAnalyse 0 : opereren op conceptueel niveau de verschillende niveau s om naar een IS/DB te kijken noodzaak tot abstraheren/conceptualiseren i.v.m. complexiteit de conceptuele informatie-processor van external via conceptueel naar relationeel niveau ORM: Object-Role Modeling De 4 niveaus van de database development cycle 1

2 The physical view Voorbeeld gehaald uit Guide to ORM (blz. 11): Sample database code (the physical view) CREATE DATABASE Info University ClassCode varchar (10) NOT NULL, Duration integer NULL, PRIMARY KEY (CourseID, Section, Semester, Year), FOREIGN KEY (CourseID) REFERENCES Course, FOREIGN KEY (Instructor) REFERENCES Instructor, FOREIGN KEY (Classroom, BuildingID) REFERENCES Room, UNIQUE (ClassCode) ) CREATE UNIQUE CLUSTERED INDEX ClassTermIdx ON ClassTerm ( CourseId, Section, Semester, Year )... It is a view that is restricted to a specific relational database and to the physical implementation of that database. In other words, the database will run only on a specific, physical data-base platform. The logical view Voorbeeld gehaald uit Guide to ORM (blz. 12): This logical view shows entities, relationships between entities, and attributes. This view is similar to the physical view; it simply replaces textual definitions with a more graphical, tabular format. 2

3 Design that communicates: the conceptual view Your goal as analyst is to drive the process of database design as much as possible from the external view. In so doing, you will produce a design that is based on sample data that users from all perspectives can understand and about which they can communicate. The eight steps of FORML methodology in the context of the four views of database design: Op de bovenste regel: Steps in object-role modeling Op de onderste regel:views of a database: Opereren op conceptueel niveau (1) 3

4 Opereren op conceptueel niveau (2) Stel je een UoD voor met het volgende conceptuele schema: Reference schemes: Person (firstname) ; City (name) ; Year (AD) Base fact types: F1 Person lives in City F2 Person was born in Year F3 Person is a brother of Person F4 Person is a parent of/is a child of Person Constraints: C1 each Person lives in some City C2 each Person lives in at most one City C3 each Person was born in at most one Year C4 no Person is a brother of itself C5 no Person is a parent of itself Derivation rules: D1 Person 1 is an uncle of Person 2 if Person 1 is a brother of some Person 3 D2 who is a parent of Person 2 nrchildren (Person) ::= count each Person 2 who is child of that Person Opereren op conceptueel niveau (3a) Reacties van de CIP (Conceptual Information Pocessor) op gebruikersinvoer: (na een accepted wordt het feit toegevoegd aan de conceptuele database) User: CIP: Add: Person Terry lives in City Seattle => accepted Add: Person Norma was born in Year 1950 => rejected ; C1 violated Add: Person Norma lives in City Brisbane => accepted Add: Person Norma was born in Year 1950 => accepted Add: Person Terry is a brother of Person Paul => rejected ; C1 violated Add: Person Paul lives in City Brisbane => accepted Add: Person Terry is a brother of Person Paul => accepted Add: Person Paul is a brother of Person Paul => rejected ; C4 violated Add: Person Terry plays Sport judo => rejected ; unknown facttype Poging tot doorvoeren verhuizing: Add: Person Terry lives in City Brisbane => rejected ; C2 violated Del: Person Terry lives in City Seattle => rejected ; C1 violated Via compound transaction is verhuizen toch mogelijk: begin Del: Person Terry lives in City Seattle Add: Person Terry lives in City Brisbane End => accepted 4

5 Opereren op conceptueel niveau (3b) Reacties van de CIP (Conceptual Information Pocessor) op verdere gebruikersinvoer: User: CIP: Add: Person Linda lives in City Brisbane => accepted Add: Person Selena lives in City Brisbane => accepted Add: Person Terry is a parent of Person Linda => accepted Add: Person Terry is a parent of Person Selena => accepted nrchildren ( Person Terry )? => 2 What Person is a brother of Person Paul? => Terry What Person lives in City Brisbane? => Norma Linda Selena Terry What Person is an uncle of Person Selena? => No Person Add: Person Paul is a brother of Person Terry => accepted What Person is an uncle of Person Selena? => Paul What Person is an aunt of Person Selena? => rejected ; unknown facttype Vragen: a) hoe zal de CIP reageren op: Toevoegen constraint C6: each Person was born in some Year =>? b) hoe kun je alsnog een conceptuele database met deze constraint erbij verkrijgen? Van external naar conceptueel naar relationeel niveau Voorbeeld van external niveau: Met bijbehorend conceptueel niveau: Reference schemes: Employee(empNr) ; EmpName( ) ; Sex (code) ; ForeignLanguage (name) Base fact types: F1 Employee has EmpName F2 Employee is of Sex F3 Employee speaks/is spoken by ForeignLanguage Constraints: C1 each Employee has exactly one EmpName C2 each Employee is of exactly one Sex C3 Sex values are M, F Met bijbehorend (logisch!) relationeel niveau: tabellen en relationeel schema 5

6 Object-Role Modeling (ORM) Š Understandable Express facts and rules in English and/or intuitive graphics Š Reliable Validate rules using English and sample populations (multiple instances) Š Expressive Capture more business rules graphically Š Stable Minimize the impact of change to models and queries ORM may be used instead of, or prior to, other approaches ORM s Conceptual Schema Design Procedure CSDP Divide the universe of discourse into manageable subsections. Apply the CSDP to each subsection. Integrate the subschemas into a global conceptual schema. For each manageably sized application,the conceptual schema design is performed in seven steps: 1. Transform familiar information examples into elementary facts, and apply quality checks. 2. Draw the fact types, and apply a population check. 3. Check for entity types that should be combined, and note an arithmetic derivations. 4. Add uniqueness constraints, and check arity of fact types. 5. Add mandatory role constraints, and check for logical derivations. 6. Add value, set comparison, and subtyping constraints. 7. Add other constraints and perform final checks. 6