Data modeling. Excerpt of the Real World. Manual/intellectual Modeling. Conceptual Schema (E/R- or UML-Schema) Transformation

Transcription

1 Data modeling Excerpt of the Real World Conceptual Schema (E/R- or UML-Schema) Manual/intellectual ll l Modeling Semi-automatic Transformation Relational XML etwork Object-orientierted Sh Schema Sh Schema Sh Schema Sh Schema 1

2 Relational a Data Model Studr Lecturer ame Students attend Title M Students Studr ame Fichte Jonas attend Studr Lecture Lecture Title r r Grundzüge Glaube und Wissen

3 Development of relational DBMS Codasyl, beginning 1960: network data model IMS, mid 1960: hierarchical data model Ted Codd, 1970: foundation relational data model System R, mid 1970: relational DBMS (research prototype) Genealogy poster: p gy 3

4 Development of relational DBMS 4

5 Development of relational DBMS Commercial relational DBMS, Focus OLTP Oracle V2, end 1970 Ingres (Berkeley), end 1970 PostgreSQL SQL/DS, beginning 1980: IBM DB2 MS SQL Server, 1990 (out of Sybase) MySQL, end 1990 since end 1990: Object relational extensions 5

6 Development of relational DBMS Specialised systems: Object orientierted DBMS (since end 1980) XML DBMS (since end 1990) Main memory DBMS (since end 1990) Analytical DBMS (Focus OLAP) (since beginning 2000) Hybrid systems: OLTP & OLAP: HyPer, SAP HAA (main memory DBMS, multi processor architecture) 6

7 Foundation relational model Let D 1, D 2,..., D n be domains Relation: R D 1 x... x D n e.g.: Phone book string x string x integer Tupel: t R e.g.: t = ( Mickey Mouse, Main Street, 4711) Schema: determines the structure of the stored data e.g.: Phone book: {[ame: string, Street: string, phone#: integer]} 7

8 Relation Phone book ame Street Phone# Mickey Mouse Main Street 4711 Minnie i Mouse Broadway Donald Duck Broadway Instance: current state of the data base Candidate Key: minimal set of attributes whose values uniquely identify a tuple Primary Key: underlined One of the candidate keys is chosen as primary key Has a special meaning when referencing tuples 8

9 University Schema require followup ame Students t attend Studr prerequisite hours Semester M M M M Grade test give Lecturer Weekly Title Persr ame Area Assistants Level work-for Professors Room 1 Persr ame 9

10 Transformation rules (1) Entity sets to relations: Entity set E with attributes t A i out of domains D i (1 i k) k-ary relation E(A 1 :D 1,..., A k :D k ). Keep key attributes 10

11 Relational depiction of Entity sets Students: {[Studr:integer, ame: string, Semester: integer]} : {[Lecturer:integer, Title: string, WeeklyHours: integer]} Professors: {[Persr:integer, ame: string, Level: string, Room: integer]} Assistants: {[Persr:integer, ame: string, Area: string]} 11

12 Relational depiction of Relationships A 11 E 1 1k 1... R 1 A A R 1 R A k R A 21 E 2... A 2k 2 E n A n1 A nk R R A11,..., A1 k, A21,..., A2,..., 1,...,, 1,..., 1 k A 2 n Ank A A n kr R:{[ ]} Key of E 1 Key of E Key of 2 E n n Attributesof R 12

13 Relationships of our example schema Studr Lecturer ame Semester Students attend M Weekly Hours Title attend (:M): {[Studr: integer, Lecturer: integer]} 13

14 Instance of Relationship attend Students attend Studr... Studr Lecturer Lecture r Studr Students attend M Lecturer 14

15 Relationships of our example schema Professors give 1 give (1:): {[Persr: integer, Lecturer: integer]} work_for (:1): {[AssistantsPersr: integer, ProfPersr: integer]} require (:M): {[Predecessor: integer, Successor: integer]} test (:M:1): {[Studr: integer, Lecturer: integer, Persr: integer, Grade: decimal]} 15

16 1:1 Relationship? Men Persr marry 1 1 Women Persr marry (1:1): {[???]} 16

17 Refined transformation rules Relationships to Relations (cont d): 1:1-Relationship: Relationship R between 2 Entities E and F. no Relation R, instead primary key of E in Relation in F or vice versa. 1:-Relationship: Relationship R between 2 Entities E and F. no Relation R, instead primary key of E in Relation F als Fremdschlüssel aufnehmen. In case R has own attributes incorporate them into F. 17

18 Refined relational schema Professors give 1 1:-Relationship Initial draft Professors : {[Persr, ame, Level, Room]} : {[Lecturer, Title, WeeklyHours]} give: {[Persr, Lecturer]} 18

19 Refined relational schema 1:-Relationship Initial draft Professors : {[Persr, ame, Level, Room]} : {[Lecturer, Title, WeeklyHours]} give: {[Persr, Lecturer]} Refinement by subsumption Professors : {[Persr, ame, Level, Room]} : {[Lecturer, Title, WeeklyHours, given_by]} Rule: Relations with the same primary key can be subsumed... but only those and no others! 19

20 Instance of Professors and Professors Persr ame Level Room Lecture Title Weekly given_by 2125 Sokrates C4 226 r Hours 2126 Russel C Grundzüge Kopernikus C Ethik Popper C Erkenntnistheorie Augustinus C Mäeutik Curie C Logik Kant C4 7 Professors give 1 20

21 Why not the other way round? Professors Lecture Title Weekly Persr ame Level Room gives r Hours 2125 Sokrates C Grundzüge Sokrates C Ethik Sokrates C Erkenntnistheorie Mäeutik Augustinus C Logik Curie C4 36?? 5052 Wissenschaftstheorie 3 Professors give 1 21

22 Consequences Anomalies Professors Persr ame Level Room gives Lecture Title Weekly r Hours 2125 Sokrates C Grundzüge Sokrates C Ethik Sokrates C Erkenntnistheorie Mäeutik Augustinus C Logik Curie C4 36?? 5052 Wissenschaftstheorie 3 Update anomaly: What if Sokrates moves? Delete anomaly: What if Glaube und Wissen is dropped? Insert anomaly: Curie is new and does not yet give lectures? 22

23 Transformation Generalization Person Birthday ame is_a Patient Insurance-r Doctor VacationDays Discussion 23

24 Relational Modelling of the Generalization Area Assistants is_ a Employees Professors Persr ame Room Level Employees: {[Persr, ame]} Professoren: {[Persr, Level, Room]} Assistants: {[Persr,Area]} Aea]} 24

25 Relational Modelling of weak Entities Students Grade 1 write Tests Part Studr consist_of give Lecturer M M Professors Persr Entity sets Students, Tests: Students: {[Studr: integer, ]} Tests: {[Studr: integer, Part: string, Grade: integer]} 25

26 Relational Modelling of weak Entities Students Grade 1 write Tests Part Studr consist_of give Lecturer M M Professors Persr RelShips consist_of and give: Globally unique primary key of the relation Pests = Studr and Part must be transfered as foreign key into the relations consist_of and give 26

27 Relational Modelling of weak Entities Students Grade 1 write Tests Part Studr consist_of give Lecturer M M Professors Persr consist_of: of: {[Studr:integer integer, Part: string, Lecturer: integer]} give: {[Studr: integer, Part: string, Persr: integer]} 27

28 Relational Modelling of weak Entities Students Grade 1 write Tests Part Studr consist_of give Lecturer M M Professors Persr RelShip write? 28

29 Professors Students Persr ame Level Room Studr ame Semester Lecture Title 2125 Sokrates C Xenokrates 18 r 2126 Russel C Jonas Grundzüge Kopernikus C Fichte Ethik Popper C Aristoxenos Erkenntnistheorie Augustinus C Schopenhauer Mäeutik Curie C Carnap Logik Kant C Theophrastos Wissenschaftstheorie attend Feuerbach Bioethik Der Wiener Kreis Studr Lecturer require 5022 Glaube und Wissen Predecessor Successor Die 3 Kritiken Assistants Persr ame Area Boss Platon Ideenlehre Aristoteles Syllogistik Wittgenstein Sprachtheorie Rhetikus Planetenbewegung test 3006 ewton Keplersche Gesetze Studr Lecturer Persr Grade 3007 Spinoza Gott und atur Weekly Hours Given_by

30 Sample assignment Assignment 1 (E/R-Modelling) 7 Points Given the following E/R-Model (in the notationof ti the lecture) with an excerpt of the world of literature. ame Writers [1,] compose Publisher [1,1] ISB Books Can a book be composed by several writers? yes no Can a writer compose several books? yes no Can a writer also compose no books? yes no Give a short explanation in each case. 30

31 Sample assignment (cont d) ame Writers [1,] compose Publisher [1,1] ISB Books Describe the two tables which stem from the refined transformation of 1: relationships to relational schemas for the above model. Use the short notation T(A, B,..) with T: Table name, A and B attributes, A primary key. 31