1. 1 Databases 2013/2014 Introduction

2. 2 The menu for today Organisational aspects Introduction to database technology The relational model

3. About using laptops in classes 3

4. Organisation Hoorcollege Donderdag a.s. in Went-blauw, eenmalig Werkcolleges (beginnen volgende week dinsdag) Huiswerkopgaven (3x) Practica Opgave 1: casusbeschrijving, modelleren, schema-ontwerp Opgave 2: vulling van de database, SQL queries Voor elke opgave twee practicumsessies op tijden werkcollege 4

5. 5 Introduction to database tech. What are databases? Relational data model (short version) Why should we look at databases? Short history of databases Some aspects of database technology Query languages Database applications: 4GL, constraints, reports ER-modeling Normalisation Transaction processing

6. 6 What are databases? (1) Example: library system Books, readers, loans, reservations Loaning books, returning books, searching, making reservations, subscribing readers BookReader Loan

7. 7 What are databases? (2) Manipulation of data using a query language For example SQL Integrated in a 4GL Often client/server architecture Application logic in the client SELECT Title FROM Book WHERE Author = ‘Kasparov’ Database server (DBMS) Client Database

8. 8 What are databases? (3) Characteristics of a database environment Stable structure of data Large volumes (external memory, persistency) Good performance More than one user at a time Reliability and integrity of data Conceptual approach

9. 9 Why look at databases? Databases are omnipresent Database technology is directly applicable Database technology is the backbone of most information systems Studying database technology provides insight in general principles of computer science Layered software architecture Application of predicate logic Mathematical modeling

10. 10 (Pre)History of databases (1) Magnetic tapes were available in 1945 searching in external memory IBM introduced the RAMAC system in 1957 hard disk In 1961, Integrated Data Store was the first general DataBase Management Systeem (DBMS)

11. 11 (Pre)History of databases (2) The Information Management System by IBM followed in the mid-sixties Hierarchical data model In the same era the Network Data model became popular (CODASYL standard) During the seventies, the commercial use of databases was rapidly growing Codd proposed the relational data model in 1970 The basis of almost all modern DBMSes

12. 12 History of databases (3) In 1976, Chen introduced the Entity- Relationship (ER) model Conceptual modeling for databases Simplified the design of databases Query languages such as SEQUEL (SQL), QBE and QUEL were designed

13. 13 History of databases (4) During the early eighties, the relational data model received widespread commercial attention In 1983, more than 100 RDBMSes existed DB2, ORACLE, SYBASE, INFORMIX, INGRES DBASE, PARADOX, MS-ACCESS SQL became a “standard” in 1986 SQL92/SQL2, SQL3, SQL2003: ANSI standards

14. 14 History of databases (5) The first 4GL languages appeared during the eighties Object-oriented databases were introduced at the end of that decade Focus shifted to extending features and better performance Multimedia databases, web databases, parallel processing Core database technology is now quite “stable” Databases + …

15. 15 Query languages From “how” to “what” SQL is declarative SELECT Name FROM Book, Loan, Reader WHERE Book.Title = ‘Fun-fishing’ AND Book.Bno = Loan.Bno AND Loan.Rno = Reader.Rno Book.Title := ‘Fun-fishing’; FIND FIRST Book USING Title; WHILE DB-Status = 0 DO BEGIN FIND FIRST Loan WITHIN Book_Loan; WHILE DB-Status = 0 DO BEGIN FIND OWNER WITHIN Reader_Loan; GET Reader; PRINT(Reader.Name); FIND NEXT Loan WITHIN Book_Loan; END; FIND NEXT Book USING Title; END

16. 16 Database applications (1) PROCEDURE Loan; begin $today := call_system(‘current_date’); read($x); # read Rno if call(Rnocheck($x)) = 0 then begin message(“ticket invalid”); exit end; read($y); # read Bno while ($y <> EndOfLoan) do begin call(Register_loan($today, $x, $y)); read($y); end end {Loan} PROCEDURE Rnocheck($x); begin SELECT COUNT (*) FROM Reader WHERE Rno = $x; end {Rnocheck} PROCEDURE Register_loan ($today, $x, $y); begin INSERT INTO Loan VALUES ($y, $x, $today, NULL); end {Register_loan}

17. 17 Database applications (2) CONSTRAINT constr1 (SELECT COUNT (*) FROM Loan WHERE Return_date IS NULL GROUP BY Rno) <= 6 ON VIOLATION … CONSTRAINT constr2 (SELECT COUNT (*) FROM Loan WHERE Return_date IS NULL GROUP BY Bno) <= 1 ON VIOLATION … CONSTRAINT constr3 (SELECT Bno FROM Loan) IS CONTAINED IN (SELECT Bno FROM Book) ON VIOLATION …

18. 18 Database applications (3) Report writing @name @address Dear mr/mrs @name, On @loan_date you have borrowed the following book from our library: @title by @author. We kindly request you to return this book as soon as possible. SELECT Name, Address, … FROM Loan, Reader, Book WHERE Loan.Rno = Reader.Rno AND Loan_date < ‘01.12.2010’ AND Return_date IS NULL

19. 19 Database applications (4) Database server (DBMS) User interface Application programs Database SQLData File access  Sophisticated user ( DataBase Administrator, developer ) SQL Queries Updates Menus Screens Reports Naive user

20. 20 ER modeling BookReader Loan (0, n)(0, m) Loan dateReturn date NameAddress RnoBno TitleAuthor Book(Bno, Author, Title) Reader(Rno, Name, Address) Loan(Bno, Rno, Loan_date, Return_date)

21. 21 Normalisation (1) Why don’t we put everything in one table? Manageability Prevent redundancy and inconsistency Adequate representation (without NULLs) Normalisation “Vertical” splitting of tables Several distinct normal forms and algorithms

22. 22 Normalisation (2)

23. 23 Transaction processing (1) Transactions are important in case of crashes and simultaneous use of the database by multiple users Transactions transform multiple database operations into one single atomic operation (as seen from the outside) Read balance accno. 1234567 Read balance accno. 7654321 Withdraw € 50,- from 1234567 Deposit € 50,- on 7654321 Write balance accno. 1234567 Write balance accno. 7654321 CRASH!

24. 24 Transaction processing (2) 1. Read balance accno. 1234567 2. Read balance accno. 1234567 1. Withdraw € 500,- from 1234567 2. Withdraw € 500,- from 1234567 1. Write balance accno. 1234567 2. Write balance accno. 1234567 Concurrency problem Solved by locking based techniques

25. 25 Why relational databases? Software Engineering High level data specification and manipulation Philosophy with regard to system development Start with rigorous design of tables Stable; detailed inventarisation is possible Development of operations is secondary Difficult to analyse, rapid prototyping, continuous adaptation Successful application of computer science Set theory, predicate logic, optimisation, design theory