Deveka:P Database Management Systems R. Shankar me, Assistant Professor and Head Department of CSE Indira Institute Of Engineering And Technology Thiruvallur — 631 203. P. Kalamani Adhiyaman Engineering College Hosur. New No. 24, Thambiah Road, yverybookpalace.comDEC. 2012 Price : Rs. 200/- ACKNOWLEDGEMENTS | sincerely thank Thiru V.G.Rajendran Chairman, Indira institute of Engineering and Technology for facilitating me to bring out this book. My wholehearted gratitude to Tmt. Indira Rajendran Chairperson, Indira Institute of Engineering and Technology for her kind Co-operation, 1 am grateful to Dr.N.Vasudevan M.Tech., Ph,D., Indira institute of Engineering and Technology for his encouraging words for book. 9 this extend my gratitude to all my departmnet colleagues for their constant support and valuable suggestions. | feel privileged to extend my deep sense of gratitude to my parents and wife who supported me fo carry on the work with their moral support and encouragement. | also take this opportunity to thank Charulatha Publications for publishing their interest and professionalism in bringing out this book. R.SHANKAR‘Buun, eseqeied - uonesodg ulor - Bumyog - uoHeIedO onoajag - uoRews3 105 40) uoneUUo}U BoleieD - maIN‘eNO Bulsseoog Kuan = Bulyse} leuk - BuIyseH ONe%S - Sell} xepUI 991) g - SAIL XOPLI 2a) + - SOOIPUL pavapio — Suse pue Surepul - sali4 Ul spi0994 Jo LoHeZIUeBIC - UONEZIUeBIO a4 — aBeu01s Avena) - Give - S¥SIG aNeUBeW) - eIpay @Bes0IS JeOISAY JO MAIAIBAO ‘S3NDINHDAL NOLWINAWATdINI- A LIND ‘Kouaunouog 404 senwtoes JOS - s19n27 uoHe|Os| Kuenoo9y - Aayigezeues - yoojpeaq — Bury907 wWayU| ~ 6urYOo7 SEUId OM - $]090}01q GuIya07 - KouasiNoUED 40} PEeN - AoUEUINDUOD ~ kuenoo—) 10} SORIIOBI TOS - SIUlod OAC - ILIOD eseYd OM - KreAoDey eIpayy - K1oRDoy waiskg - somedorg GIOW - fuenovey UoNesuel) - sideouog uoHesueLL SNOLLOVSNVYL- Al LINT uo [@WON Yl pue se!ouepuaded Wor ~ woy JeWON ynoy pue se}ouapUedeg panjen-nINW - WO JeULON PpODeoKOG = uonenioselg fouspuadeq ‘suo jeUON PUIYL "PUodag Ysu1,4 - saloUpUadeq Jeuoyouny — uonisodwoseq ssoj-uoN - seisuepuedeq jeuonouny NoISaq aSvavivG - II LIND. ‘soseqejeg Jeniagiualia pue saseqeyed paynquisic o} ‘uononposquy ~ smeiA - UoReLUUOJU} BUISSIW - TOS DIUHeUAA - 1S peppaqui - saimea} OS poouerpy - Auunsag - s1e66u) - AwBeru) - sjeTusWepUNJ TOS - SuoNeIedo, euonippy — suonelado jejuewepun, - snjnoje9 jeuoneiay ejdn4 - snjnsjeo [euoNe|ey Uujewiog ~ eigaBjy jeuonejoy - SKoy - sadK| — Bojejeo at - japo\ jeUOHe|al SUL JaGOW WNOLW 134 - It LIND soseqejep leuoyelas 0} uononposiul- swes6eIG Y-3 - (Jepow Y-3) Jepow diysuoNejoy - ANUS = soyes)siulwupy pue siasn aseqeleg - esmppoyyosy Waisks aseqeieq ~ saBenbuey ASEqEIEG - SIBPON BIE - BIE J0 SMaIA - WIa}SKS aseqe}eG Jo osoding NoWonGowLNt - LINN (18 3S9 0) voWWOD) ‘SWALSAS LNAWSOVNVIN SSVEVIVG - 992289 siouiny ‘sjuowonosdu J0ULM} 40} paBpeymoUrse Aiinjayes6 2q j1Mm sjuouMOD pUE suoNsaBEng (g002 - e002) siadeg uonseny ysinlun eUUY PeNOsUN pUE 4007 uOHeIBey - sieded UotsenO Kisterlny ey Beniog ‘Jun Yoke 40) ¥y = ed - uor|sano meinay Paniog sopinoid 4009 oyy 19 Hed pu y— Hed 10) Uan6 ase suonseng) mainoy wun Yoes aun Jo pue oun y ‘uBisep eseqejep jeoishyd uy djoy ue aiy ueveyIp Bursn aunyonyjs oBeio%g jewoqU) au U {yoiym sanbiuyoa} Burxepu} sleap g- Wun ou, Ionuo9 fovexmoueg ou pueKionoe ‘Buisseocig uoNesueL| yo sjdaou02 TeIUSWEPUNY 84) YIMM Sjeep y - Wun ayy s *0004d vonezIeWoN pue s;LopuedopfevoHauN §NOGE SessNosip Ose y “UB|SEp aseqeyep jeuoNe|a, a) Uo S@SNO0) ¢ - HUA oy rans soseans J2viegnuot pue seseqeiep parnaiiep 0} uononponuy pue uonewows ares “Aunoos pus fab rage soup one Se suepung os pe ebccbee at Sedhy "apou uonejes 91 Uahanpoiul uy ln Sopsoe fe en "8002, uoje;nBa1 10j sngeys pasinas ysore| YM AySsenlUn euL 49 swopmis (11) Yoel @ 389) “Ja au J0) UHM uoag sey yoog siy] 3OVdadd2.3.2 Network Data Model 23 CONTENTS 2.3.3 Relational Data Model 24 UNIT 1 2.4 Physical Data Models 25 CHAPTER -1 Relati INTRODUCTION TO FILE AND DATABASE SYSTEMS, 2.85 . Entity Relationship Model 4.4. Introduction fa 2.5.1. Basic Concepts “ 1.2 Database System Applications 42 2.5.2 Constraints an 1.3. Purpose of Database System ag 2.5.3 Weak Entity Set . an 14 Views of Data a 2.5.4 Strong Entity Set z an 14.1 Data Abstraction ae 25.5 ER Diagram 2.16 1.4.2 Three Level Architecture ae 2.6 Introduction to Relational Databases 2.46 1.4.3 Instances and Schemas an 2.6.1. An Informal look at the Relational Model 1.5 DBMS Languages at Review Questions 221 1.5.1. Data-Definition L: anguage 442 ce 1.52 Data-Manipulation Language . a 2 CHAPTER -3 1.6 Database system architecture 413 _ INTRODUCTION TO RELATIONAL MODEL 1.7 Database Administrator ial 7 1.8 Advantages of DBMS 116 3.1. Structure of Relational Databases 34 1.9 Disadvantages of DBMS ne 3.1.4 Basic Structure i 2 Review Questions 48 3.4.2 Database Schema 3 3.1.3 The Catalog 34 CHAPTER -2 DATA MODELS 3.1.4 Keys 37 2.4 Introduction a4 3.2 Base Tables and Views 37 2.2. Object-based logical models, Pei 3.3 Transactions 2.3. Record Based Logical Models 22 3.4 The supplier - and - Parts Database 38 2.3.1 Hierarchical Data Model 22 Review Questions 3.1069 eg 619 ero a9 sto eg eo 9 ore 69 89 #9 “9 “9 99 99 99 eg zo zo bo rz) suoisuayxg age AUD 4279 saunjea TOS PesUeNPY 7Z'9 Jos oueuka 129 Tos Peppsqui3 029 suior 619 ‘aBenBue7 jonuoo uonoesues| g1'9 smeI L1'9 seuiand xeiduiog 919 souianbgns paiseN SL'9 SenIeA INN ¥L'9 suoyouny oeei6By €1'9 suoneiedg es Z1'9 asneig Aq s8p10 119 suoneiedo Buns 01'9 sajqeuen aidnt 69 uoneiedo aueusy 99 uoissaidx3 TOs Jo einjonysoiseg 19 eGen6ue7 uoneinduen Bea 99 aGenBue7 uonuyed Bed So Jos ul sedk uewog 79 qosso sued 9 os jo sebewerpy 79 ononponuy 19 STVLNAINVONNS TOS 9- YALdVHO as sis es es zs as us Ws Ws Ws as gs VS vs bs ey vy ey zy by Suonsany memay snjnoje9 jeuoneloy wrewog ay, —2'F'S sveyquend jesienun pue jequelsiKa ZL "b's SPINWUO. PUR SUOISSAIGX || -p'g shinojeg ieuoneey edn, ay, yg sninojeg jeuoneiey yg uonepdn ge's uomesul z'¢° uonaiea | Ves aseqeep om Jo uoneoWpoW Eg suoneiedo eiqaBiy-ieuoneiay popuaxg —¢'2's suoneadg jeuomppy zz suoneiedo oseg 1'2'g eugaBiy leuoneiey zg uoyonpoauy 4'¢ BOVNONVT ANAND $- MaLdVHO suoysand money somes tos yy soyeioueg adh, e'y SOIGeUEA SA SONIEA ZY uononponuy 4" S3dAL ¥~YaALdVHO6.22.2 More on Subqueries 6.22.3 Advanced Constructs for Database Update Review Questions CHAPTER -7 INTEGRITY AND SECURITY 7A 72 73 74 75 76 7 78 79 Integrity Constraints 7.1.1 Domain Integrity Constraints 7.1.2 — Entity Integrity Constraints 7.1.3. Referential Integrity Assertions Triggers Security ‘Authorization Authorization and Views Granting of Privileges Audits Trials Security Specification in SQL 79.4 Pri leges in SQL 792 Roles 7.9.3 Limitations of SQL Authorization 7.10. Eneryption 7a Authentication Review Questions 6.20 6.21 6.23 7 7A 72 72 74 74 76 7 78 79 7.10 7.40 7.10 7.1 m1 712 7.412 7.14 CHAPTER -8 MISSING INFORMATION 8.1 Introduction 8.2 An Overview of the 3VL Approach 8.3 Some consequences of the foregoing scheme 8.4 Nulls and Keys 8.5 SQL Fa Review Questions ties CHAPTER -9 INTRODUCTION TO DISTRIBUTED DATABASES AND CLIENT/SERVER DATABASES 9.1 Introduction to Distributed Databases 9.2 Homogeneous Distributed Databases 9.2.1 Data Replication 9.2.2 Data Fragmentation 9.2.3 Transparency 9.3 Heterogeneous Distributed Databases 9.3.1 Query Processing 9.4 Client - Server Systems Review Questions UNIT- CHAPTER - 10 RELATIONAL DATABASE DESIGN 10.1 Introduction 10.2 Functional Dependencies 10.3. Functional Dependency Theory a4 82 86 att 815 a4 93 93 o4 95 96 97 97 99 10.4 10.2 104eezh bea we vee web ozzh ebzb wat Wb over Orel 6zl 8th oz val va ba zee one ore eu ou suonsand) MAIReY, somyges OS LL proy Budo oF Zt Suppor wey} ZL sens] uonejos| e'Zh poysinoy Aionovey L-Zb Augezeveg 9°21 spoypeed 10} Asnovey Z'¢'S'Zh uonoaieg yoIpeed 1'E'S'ZL ‘Arsnooey pue uoy22}9q Y90}PeeG E'S'ZL sewaypg paseg-ynoeull| ZZ woquandie Y9O}PEAC 1'S'ZL BuypueH yoojpesd Ss zh Jooo}o1d Bupyoor aseyd -OML YZ Supo7 e'Zh Aouaunouog 40} pPeeN ZZb uononpomul 12h AONAYYNONOD %b- WBLEVHO Suonsanp mained ru ou yuwiog eseyd-omL Lh Aronoooy worskg yl eu zu ba ozor stor 2vOr vol Vor svol — fusnooey uopoesues) €"4 soqeig uonoesuedL Z'L1 yde0uc9 uoyoesues] 0} UoHONPONU] L'bL XSBAOOTY NV S1d3ONOD ONISSIDOUd NOILOVSNVUL bE Mad VHO AV LINA, suoyseny mainey (4NDIA) wH9y jeuOU fay UEWOg ZO} wo} jeuOU Yi pue sé uepuedap ulop OL SUONEIaY Ny O}U! UOH|soduiod8q ulor ssa} $501 ¥'¢'O} Wo FEUNON YUNOY €°S°O} ‘Sejouepuadeg penjenninwy pue jeuoyoUn 404 Sajny eoUarOyUI Z'G-O1 svor svor yLoL yboL ror eor wor eo. Zor gor sol fouspuodeg penjenninyy Jo u uyog JeUtioy 1'S'OL ‘oy JEWON YpNo} pue sajouapuadep panjeaINN s'OL LANE PUE INDE JO UosHEdWog g'yOL sulop ssaisso] ZS OL uonenteseig Aouapuadag | 'S'Y'OL Uuonjsodwoseq uoneay ¢'y'0L 04 [WON ppoo-eofog y'y'OL (4Ne) WuO4 JEWUON py, EP OL (4NZ) wos JeuuoN puosdsg Z'y' OL (AND) wHo4 fewuon sid OL UOHEZIIEWON OL| ante 13.8.2 Vatiable-Length Records 13.16 HAPTER -13 18 rgaization of Records in Fes 13.17 ee ee eerie 13.9.4 pee Unordered Records (Heap Files) a) | a ee eee tat 43.9.2 Sorted Files (Files of Ordered Records) 19 | a2" Primary SoregeDe 13.9.3 Hashing Techniques 21 2. Primar mary Storage Devines ue) 13.9.3.1 Internal Hashing for Main Memory 13.21 | 13.3 Flash Memory 13.3 13.9.3.2 External Hashing for Disk Files 1323 | 13.4 Magnetic Disk 13.3 43.9.3.3 Hashing techniques that allow dynamic fle expansion 13.24 13.4.1 Introduction 13.3 Review Questions 13.30 13.4.2 Magnetic Hard Disk Mechanism 13.3 CHAPTER - 14 13.4.3 Performance Measures of Disks 135 noeane! 13.4.4 Optimization of Disk-Block Access 136 44.1. Basic concepts wt} 135. RAID - (Redundant Arrays of Independent Disks) 137 142 Single level ordered Indices 142 13.5.1 Improvement of Reliabilly via Redundancy a7 | 142.1. Primary Index , 442 i 13.5.2 Improvement in Performance Via Parall 138 1422 Clustering index 143 1355: ocr 138 44.2.3 Secondary Index 144 13.5.4 Choice of RAID Level 13.11 | 443. Mullievel indexes 145 13.6.5 Hardware | 7 | tN) 13.11 44.4. Dynamic multilevel indexes using b* trees and b-Trees 147 13.6 Tertiary Stor ry Storege 13.12 14.4.1 Search Trees 147 19.8.1 Optical Deke ‘32 1442 B-Trees 148 13.6.2 Magnetic Tapes (coming under tertiary storage) 13.13 | WAG epee 14.10 | 187 Storage Access 13.13 tart Botorw 14,4.3.4 Observations about B'-Tree 14.12 uffer Manager 13.14 14.4.3.2 Queries on BTree 14.12 13.72 Buffer-Replacement P 13.44 14.4.3.3 Updates on B'-Tree 14.13 13.8. File Organization 13.18 14.4.3.4 Br-Tree File Organization 14.16 13.8.1 Fixed Length Records 13.15 ‘ Review Questions 14.18seb ben ee} vest zsh oz'st erst rst aust vist vest vest 8002 UoneinBay - siodeg uonsend Ausionlun euLry Pens 002 uoneInBay - siedeg uoRSeND AysienUN euUry PeNosUN V- Hed - siomsuy’g suionsenD malay suogsanty Maley soujeping Buluny Aieno jeuomppy F'8’St soyeng Buuny e's'sb ubisaq eseqzieg emp Buluny z'e'St sexepuj Bujunt 18'S} suiaiskg jeuonejoy ul Bujuny aseqeyeg Jomamniong uy 8'Sh Buyedig TLS) best suojssoudxq Jo uonenyeag 1'St uoneZIeLaIeW a erst erst erst erst eLst eLsh zsh zest zsh West ost ost 6st est est est scr vst est est esb zsh vst ur 11no '9SL ( paw un) suonesado tes ya'st voneBaibBy ¢'9'S1 uonoaloig z'9Sh on eyeoydng 19st suoneiedg 1010 suiop xejdwiog ¥'¢'S'S} uiop-yseH pughH €°S°s'St wop-yseH Jo 1809 Z°S'S°SL ‘smoiieno BuypueH 1's°S°S} wor-useH $'s°S1 wop-26eN ¥'9'Sh or door - parse pexepul ¢'S's1 lop dooy-poysony 001g z'S'SL op door-paisen b's'Sb uoneiedg wor Bumog ySL suojosjag xe|dwog Jo uonejuoWe|dw] p'E°SL suosuedwog Buinonuy uonpaag €'¢"S1 seolpul Buisn uonsajeg Z'e"Sb (yoieeg seour) yy uIHOBIY L'e"St uoneredo uonoejag 1809 Aten Jo sonseay Manian 9st ssh vst est zsh rsh ONINNL 3SVEVLVG GNV ONISS39OUd ANAND SL - YaLdVHO‘Unir-1 INTRODUCTION Cuarter - 1 INTRODUCTION TO FILE AND DaTaBaSE SYSTEMS BBEBBBBBB Introduction Database System Applications Purpose of Database System View of Data DBMS Languages Database System Architecture Database Administrator Advantages of DBMS Disadvantages of DBMS“ soquep ay ur poureiuod uoretuzoyuT axp Burney pue Busors| “Buruyep yo powpauu aandays pure woruaatoo apyaoxd 011 SIC tn 0 [08 a, SWAJ0 S10 “asequrep B Ur poure|uoo eyEp stp ssa0oe stosn 94} 0} SMO|TE IBID ABEANFOS St SEI Go) ‘aseqerep & ureyueus pure ayaa oy sx050 soqqeu you stresoud yo uor}]00 vs (SHY) wIsKS woUBdoUNY aseqEIEC ¥ swaa 00. | sem | go0r ‘sajdnj (10) spuooai (10) sor <————<] ggg | wefexg | zoo 00s | poeuy | roor somqime (10) spRay (0) suumjoo <—_——[ Kroes | aureus | ous ure ayy (20) owe 9pqe. <———— agoydurg, “Saoqes paw 110} JO sissuoD esequIup 2aXojduro s9pisuod opdumexa 104, “saqiy ayn wasasdar soyqea pure sprooas ayp ywosoudas swvor ‘spay arp quasaudan suumyoo aroun ‘S9jqei Se pazruedu0 won eausoyUr payp|a4 Jo wonDdITOo w st aseqeIEC asequied, ‘sossauppe pu Satu ‘ou jor yuopmg ayduexg, -Buqueote yforjduu axey yeyp pur paprosar oq eo Ip SIAR UAOUS, ued NOWONGOMLNI FF SWALSAS ASVAVIV NV TTA OL NOLLOAGOULNI 6 waLdVHO12 Introduction to File and Database Systems Database Management Systems Database System ‘A database system is an integrated collection of related files along with integration of the data contained there in. Users/Programmers DATABASE ¥ SYSTEM ‘Application Programs/Queries ams v SOFTWARE ¥ Software to Access Stored Data ZN Stored Database felis (Meta-Data) Fig. 1.1 Simplified Database System Environment 1.2. DATABASE SYSTEM APPLICATIONS Database system are widely used and the various ap 1. Banking- storing customer information, aecoun 2. Airlines-to store reservation and scheduling information. 3. Universities- storing student information and course registration. | @._ Finance- for storing informat 5. holding, Sales, ete munication: for keeping records of calls made, gener 6. Human resources- for information about employees, salaries and payroll taxes, etc. 4.3. PURPOSE OF DATABASE SYSTEM To understand the purpose of database sy enterprise that keeps information abot us consider part a savings-bank ‘manipulate the fl & A program to debit or credit an account ® A program to add a new account + Aprogram to find the balance of an account * situations that do not arise in savings accounts, such as overdrafts. Thus, as time goes by, the system acquires more files and more appli This typical file-processing system ‘The system stores permanent record programs to extract records from, management systems (DBMS) came along, organizations usually stored information in such systems Disadvantages of File Systems Over Database Systems (1) Data redundancy and Inconsistency (Data redundaney means same info For exampl appear ina file checking-acco' may be duplicated in several files. \e address and telephone number of a particular customer may ‘consists of savings-account records and a file that consists of ecords in a bank, This redundancy leads higher storage and access cost.‘eU9ypS peuUsd Xo 1 Jo suwau £q paquosap st nate qeuiarxe yor “papnyour axe suressoud us Jo asequiep sip Jo suomiod asoup Kuo aioy uorroessge asequrep Jo JaA2} 3S9y3ty Oy St tA 20 peusoyxg, soy weyd pur spiogar 9} Yee We uiaisfs @ suvaut ssaquosap Temp UL Ho] 40 pempdaouo9 ap ‘Jo4o] sata som sxoumRy) fopow (2anftuUO> squoWEHINbIY, wy) DUVAS/ISNV UL pow oq yous upjd v so autjino ure s} auiayps y 2uH94> aysiyd 20 peu) up :speaa] 2am pur Suruuvjg spimpueig/aininsuy sprepueig jeuOHTEN eamppaqyouy Jno} BQUL Z'PL ‘ayn ypym suonoesaqu s,rasn ‘Autwajduuoo otf apiy siadojaxsp “pouyen soyndusoo you are s: Douig “asvqniep 4p Uy eIep ywasasdas o4 Sammons sn suaysés-osequiep Sueur texep xojdusoo asn 0} siouasop pal Set ‘a|qussn 9q 01 warss oup 04 uonoensqy 8G 1b -pourejurew pure pasoys axe vyep axp moy Jo s|eoiap wrer.26 sopry woyss ayy ‘S! HL "wIEP aH {Jo sata jaensqe te yptm Stasn aprAoid 0} sf onsks aseqerep vJo asodind sofeu y VIVG 40 SMAIA vb padojanap 1p 9804p) BUODIOAO OF som (SWAG) warsKs iuowEseUELU esuqEIEP aKp swuowieBeuew oup Jo s|ibiep a4) ss990e you INQ 8 12S) “UIUpe 49sn pure “es ost ‘surei@oud wonwoyydde wuar2ysip Aue asneag aprscud vuorsyadns jo wioyy awos ureyureut smu woysKs ‘sity plone oy, ‘suonrenys Aupus ur ‘sis 9seqeEC Pu IL 0 OONpKLio Introduction to File and Database Systems ‘STUDENTS. NAME O° [STUDENTS ADDRESS ‘STUDENTS.FEE_DUE Logical Record 1 STUDENTS NAME STUDENTS.NAME STUDENTS.LOGINNO_ | ©, STUDENTS. COURSE STUDENTS MARKS Logical Record 2 STUDENTS.LOGIN_NO © [ STUDENTS. ADDRESS ‘STUDENTS COURSE ‘STUDENTS FEE_DUE ‘STUDENTS REGIN_DT DBA ‘STUDENTS MARKS ‘Conceptual Record Fig. 1.8 Conceptual record and two views on it ‘The data can be abstracted in three levels corresponding to three views as shown in 1.4.3 Instances and Schemas Database change over time a’ informat of information stored in the database at a particu database. The overa changed infrequently if at all Database systems have several schemas, ‘The physical schema describes the database desig schema describes the database design 1. Adatabase may also jemas, which deseribe different views of the database. Uni the structure ofa database is the data model: a co of conceptual Database Management Systems Lageatsiee ‘STUDENTS.NAME_ “| /ENTS.NAME e STUDENTS.ADDRESS STUDENTS.LOGIN_NO- STUDENTS FEE_DUE STUDENTS COURSE Logical Record 1 STUDENTS MARKS User! Logit Record2 STUDENTS NAME : STRING STUDENTS LOGIN_NO: Pi ‘STUDENTS ADDRESS : STRING AARY KEY Conceptual view ‘STUDENTS COURSE = STRING )ENTS.FEE_DUE : INTEGER STUDENTS.REGIN DT: DATE ‘STUDENTS. MARKS : NUMBER Conceptual Record DBA SO. [ NAME: String length 25 ofset 0 Internal view LOGIN NO: Sting length 10 offset 25 "ADDRESS : String length 30 offse 35 COURSE: 51 [FEE_DUE: Integer length 4 offset 75 REGIN_DT : Date length 9 offet 79 MARKS : Number length 4 offer 88 Internal Regord data Fig. 1.9 Three views o” 1.5 DBMS LANGUAGES A database system provides two differe 1. Data - definition language - To specify the database schema ‘manipulation User2 es of languages. They are nguage - To express database queries and updates.°s, uyopUI asoKp 21035 0} pasn sauManIIs BIE pure ystp wo s0eds aBes015 5 sofeuour pups saSounpy apy (¢) Supaqsu0d mowpim pasooud suonnoax9 re saumyey wonshs 4p way ares warstsuo> sainsua yoIys s28ounyy uonoesunss (C) "ep ssa908 0} szasn Jo AvLoYN ain SxDayO pu sHUMEASTOD AsBarU Jo uonoRystES ax 10 sSOi YIN :zaSounpy MsFaIU] pun uoywroWMY (1) “squauoduroo Bumo}}oy 2 Sopny -aseqenep aip ut evep ayn Bunepdn pur Surxoinay “uyors 10} a[qisuodsat st JoBeuews aBeso1s aqp SMU Aly FAA] MOT OWT SHUBWANES "NC snoLWA axp sarejsuEN soBeuEUT aFEIONS ot, “uorsis ax 01 Ins souonb pure suressioxd idv xp pre aseqerep axp tt pasons BIE [SAD] MOT aN HaDNIq DORPLOTUT raoud yet o|npout urexBoud w staf “aoeds afer07s Jo wunowre are] ve sasinbar Aipeardss asoqerep asneaoq weniodun st 4 :2aSmumpy 280015 (0) srovensjutwupy pure siasp asequieg (0) “ossaoosg Azand (4) ‘raBeunyy aBns015 (6) :syqun ¢ mut popratp ag to uraysks asequrep & Jo swuauodwoo jeuonoUTy aU, BUNLOLIHOUY WIISAS SSVEVIVO 9b “wep sor 198 01 Moy SurAyFo0ds anos popoatt are Ep TysA{ytoods oF Josn w aumMbay STING [eAnpa201g-10N ‘rep 96017 198 01 Mot] pu papaou are MEP rey Aj!2ads 01 Josn w aun STING Teanpooord - wanpaaoud-von, eanpa0c.g NC uoHeWOgUI Jo suoneoyrpour yy < uoneUOyH Jo UONOTP aL —_-< rt ‘SUDISKG wowsdeuayy oseqEIeG uoneunojyy sou Jouomosu ay, < 1ui0yu J0 [Pas apnjauy suo aseqerep at s asoquiep amp JOS} £q wiorsts asaqurep aup Aq pasn sp sso9on pup amonuis aBero1s ams oiendury] uopuyog pur a3ei0.5 Wea jp afiesn samnosey aseqpnyp ayy o1uy pode a1 yap sammonus ofBo} Jo sy < usojutyontu0> ssa00y saunyvay uisop pue swarsss Buypsefos uyaoy puv oyp 199491 0% ‘soBueyp tno Soure9 VAC OU eSunnaaxa £q wwayos asequiep ‘apnyout yc 8 Jo suonouny au, (Vg) sovenstunwupe ‘aseqetep v payfe> si wiorsks au s9n0 jontioD jenuaa Yons sey oy ostad Y HOLVULSININGY aSvevIva Lh “Suaisiy asequieg pue ajid OF woHTDNPOZIU) vlIntroduction to File and Database Systems REVIEW QUESTIONS PART-A A briefly the various levels of data abstraction. ‘ov" 2002 & May 2007] mush between physical and logical data independence. [May 2003, Nov’ 2006 & Nov’ 2007] 3. What is a data dictionary? What are the information stored in the data dictionary? [May 2003, May 2004, Nov’ 2006 & Nov’ 2007] 4. What for logical level is used in the data abstr [Dec'2003] 5. Define schema and instance. [May 2004 & Nov’ 2005] 6, Define database management system. [May 2005 & May 2003] 7. Define data abstraction. {May 2005] 8, List the function of database administrator [May 2005 & Dec’ 2005] 9. Define data manipulation language. [Nov'2006] 10. What are the problems in data redundancy? [May 2005] 11, What is the purpose of buffer manager? [May 2005] 12, Define the term data — independence. [May 2006] 13, Define Meta-data [May 2007] 14, Define Subschema, [May 2007] 15, What are the major disadvantage of processing system? [Nov'2007] 16. List five responsibilities of the DB manager. [May 2007] 17. List any two advantages of database systems. [Nov' 2007] PART-B 1, Explain in detail the major disadvantages of ke a file ~ processing system, ‘organizational information in {8] [Nov’ 2002, Nov’ 2005, May 2006 & May 2007] 2, What are the components used in storage manager of a database system? What are their tasks. [8] [Nov’ 2003] Database Management Systems ri Database Management Systems |. List out the functions of a DBA. jon languages are different from conventional in [8] [Nov Tn what way data ~ mani procedural languages? Exy . With the help of a block diagram explain the basic architecture of @ database management system. [8] [Nov’ 2006, May 2007, Nov’ 2004 & May 2005] Whi .guish between external, conceptual and internal schema. Explain the data manipulation language and data de: meant by aschema? Explain the data, mary in d . Draw the three schema architecture of database systems and explain in detail. [10] [May 2003, Dec'2006, & May 2006) Compare logical data independence and physical data independence. 6] [May 2003] ate database system structure with suitable block diagram. [May 2003, Nov’ 2007 & Nov’ 2006] [Nov’ 2006 & May 2008]saseqeieq [euonejey 0} Uoj}ONpoNU] repow diysuonejey-ANU3 Slepow eveq 1eoIshyg sjopow je21607 peseg piooey ‘sjepoy [201607 paseq-19lqQ uuonanposuy sTaqoy] VIV, 7 ~ MalavH, g@aqgqgaaCHAPTER e DATA MODELS 2.4 INTRODUCTION Data Model is a set of concepts that can be used to describe the structure of a database. (or) A collection of conceptual tools for describing data, data relationships, data semantics and consistency constraints. Data Models can be classified into three categories 1, Object-based logical model 2. Record-based logical model 3. Physical model 2.2 OBJECT-BASED LOGICAL MODELS lisa’ mn of conceptual tools for descri constraints, ‘There are various object-based models, > > Binary Model > Info-logical Model > Semantic Model ‘Among these models, the most widely used is En jonship model. Advantages 1. Exceptional concept ‘Visual representation 2 3. _Bifective communication tool 4. Integrated with the relational database model, ‘+g ig ut poroidap st ajdurex9 205 fopow fom uaanag diystuonjan 1oautp v 1wasoudas 07 pasn st ad ‘ap yuasaidar yin swan eyep Jo Jaquunu e jo dn apets st ‘2d dau uw ruasasdas 01 ps ‘dt plooas y dt 9s pu adKy puooar éyoureu ‘Sanna uzersioq sdiysuonefas pur Sa ‘soqeep ay) wasaxday 0} SauMjonUys WIEp TUAIEY]IP OM S9sM [OPOUL BIEP ONIN lepow eG HOMION Z'EZ spmpues Jo Ye onentouoydum, ‘ouspuadaput yoamaniys JO 997] oueus 0} 1001 wonewuatuayduy xojdwwog —*T sofemeapest ‘Koustouyt AquBoquy aseqnne ‘souopusdapuy ee, Ayunaag aseqeied Aondung yemdsouoy —“T sofityunapy repowy PIMC wwomauEsatHy 77 Bly 00! | preg | dary | at (ep fy vols] =e] © | [only oom [ upiioig ya re By Qe] 2p waWOLSND wall bm b® “CC Big uy umoys st fdurexe 205 japour jeoxyoresaHH “sad prooos ywapuadep si Jo j1e pue adfi prooes yoos » squasaidas Ktoanoaye aan yoeg “SadAi proses wasaidax j09 es asequiep ayy pu manus 2an, ‘bul palapro Jo Aes [eonyomesoy w Ut paztuEAIO s} Bp 2m ‘jopow yeoryoxeoy w uy -wrayscs pase ayy ayp toss poxjon9 fapous Twomey 2, JepoW BEA TEOSIBION }'e°2 Tepoyy feuONEIPY TPO HOMIN PPPOW TPOKSUBIOTH “21a sjepour paseg-puosas Jo sad 2anyp axe 22244], “SWEC. ‘8 30 sanbiujoay ssoooe pu saunjanuis eiep atp S2qu0sap opou paseq pu0sd4 2H, STAGON 1V9ID07 Gasva GHODsH eZ qua|uoD woNBULOpUH JO SST“ [om] xa [v [ee | 10 [owe] an Jo [sms [so] [me] a |v [ey | 19 adendury] wonepndrwes wep ON“ uwonejuasaudas diysuopoyay pouwar] °% ws] =e [0 [= [ola] [efspo]n copii aE PAR“ sofeyueapesia, 7 suaiekg wowBeunyy as6qed SPO eG Tt24 Data Models J Database Management Systems 25 Customer Sale Gorm | Name [cut [Om | [Custno] tem | | Category } Country 7 0 fof ao at r a fu | 300 “ " = cr | Alicia a UK a n | 300 2 | Malcom) A usA : 200 3 | Frans | C UK a= Fig, 2.3 Network data model Advantages 1. Conceptual Simplicity Promotes database integrity 2 3. 4 5, Data independence 6: Conformance to standards Disadvantages 1. System complexity 2. Lack of structural independence 2.3.3. Relational Data Model ‘A relation may be visualized as a named table. Each column of the able corresponds to anattribute of the relation and is named, Rows of the relation are referred to as tuples of the relation and the columns are its attributes. The values for an attribute or a column are drawn from a set of values known as domain, Relational data model inFig.24.¢ example is shown Colour Advantages L 2. 3 4 5. Fig, 24 Relational Data Model Structural independence Improved conceptual Easier database design, implementation, management, and use ‘Adhoe query capabil ‘A powerful database management sy Disadvantages 1 2. 3. ‘Substantial hardware and system software overhead Poor design and implementation May promote “islands of infor problems, 2.4 PHYSICAL DATA MODELS It is used to describe data at the low: To construct the physical data model- ‘was used, ernal level or physical level) 1d frame-memory model‘zawoysno v Jo yig-Jo-arep aangume ‘amp toy paayap st 9s Aus zauoysno v Jo ae otf ‘apdurexd 30 soanqune pareyes soyjo wioyy pantsop ste oynque Uy 22 Ww “auoyd jo suaquimu yuarayyip oney Avut ‘ssowioysno juasayip pur sou uot yexan9s 30 au0 ‘012 axe4 Kou JoWOIsTO ‘you auoued ainqumye aap yn 198 Aue zowtorsn9 ap s9pIstOD) zaydearex a “foo seynonued ‘4oy anjea auo ueyp azour Buyany angie UY 237m my (a) ‘saynqunre ponyea 2pSus axe yaans ouupu ‘Aqua sauorsno vu :opdrexg, “fas spjnorued w ut nea 2uo fyuo Buyxey aangunre wy sazngqan anyos 273 “oureu-ysoy ‘oureu-oppprus ‘2ureu- popratp aq so4any uso awe aynqune ayy “Kunua souorsno w uy :9ydUHE “sumdgns Joos owt paplaipaq ueo wep ainquue wy :aymqusam ansodway (1) setuoysno Jo prsoworsn :ajdurexg -syedgns saxpiny owt papysip aq owe rexy aingumne uy :aynquann apd: “sadfy snoyen ont pay 19 94.9 an -soingune oxp axe wos ‘aureu ‘pl sowoysno AijuD JOWOISNO op UL sonqume patje> ame Aanua ue aquosap rip saquadosd a4, sainqunv(q) sauorsn 19s Anta 2tp Jo woIsuanxa atp ue ssoUoIsNO UE TENPIAIPEL ep TTY 9 JENPLAIpUE aU, 100 8 st 998 AyUD UY 4a Aaya ono wogsuayxa 2x pore axe 79s Ato we -sonuodoad oures ain Sutey sonniua Jo woH ‘uosiod yeu) SOnnUep ‘Ajanbyun anpea asoyn £uadord pr-uosiad v oxey & Kou wostod afdurexe 104 “Aqua ue Apnuop! jonbrun {ew sopiadod jo 198 2wos 105 ssnyea sip pu ‘sarodosd Jo 9s sey ANUS UY “ayo ‘asnoy ‘9 “uosied zejnoned v :aydanexg, “sypalqo satpo [Te Woy aqqeysinunsyp st ep plow eas put ,320%40, 30, Su, vst AUD UY AS MK tt suioyeky wouRBeunyy aque :sKoydwwa jopous ya 1p sydeouog oseg 1S, up.Siyp dhysuonoy ~ somquny C — Anug =) waWOIsNd “UNGWAIHS aNVN ‘3000 oa ‘3009 as sn rat veal 6-7 Brg Uy wos sv uresBeyp (gra) drysuoneyax-Aaque Bsn ‘uBisap asoquiep 40y pre we sty < ‘SIWAICT Are Jo wuapuadapur paajoxs oq wo JOPOW < ‘poyuasaudor ome jopour y-—$ Relationships between 3 entity sts are called ternary rel Prescription Customer Patient Doctor base Management Systems 29 62 Constraints ‘An E-R enterprise schema may define certain constraints to which the constraints to which the contents of a database must conform. ‘Mapping eardi which another entity can be associated via a relations defined as the number entities to set. Fora binary’ must be one of t @ One-to-one (1: 1) associated with at most one entity in B, and an entity in B is at most one entity in A.1s diysuonejoa e ut Ainua ue jo uonediotyed jeioy wuasaidas oy — out qnoq tas Aan eam yuosasdas 1 — 9fSteio01 jqnog, panyap wasaxdas or — sasdiqja pause. saonqume panjeanjnuy wasardas oy — —sasdiye ajqnog. s39s diysuoneyas 01 sios Ayo pu sas Sano 0 soinqumne yu or — sour] sas drysuoneyar wasardas oy — spuowreig ainqune ue yuasaudas oy — sosdita 15 Aan we wuasardas oy — arBuerey uresfieqp oup ur pasn syoquass cuesBep Wa sv pyre st asequp & Jo aimonns [eo{o} e940 ayp Jo worREuasaudat jorBoy ayy, wesBelg ya $97 uaowreg addy diysuonvjos Nii v “YOI SHYOM 87'S “puot si pue xoq sepnBueyoa1 fq suiesBeyp y-q Ul parwoxpuy st 198 ayduexa 104 1p & Aq diysuoneyas nue Suons y ‘sdf Aynua Suons paleo az tino stomp Jo soinqumne Koy ancy rep sad Kinu 3S Anug Buons y'sZ “qmo1 st LNAGNadaA Jo veyp swaraym ‘Jemed 81 JAAOTAWA Jo onPdionsed ou, “LNAGNadaTG Ad4 An years auf 405 diysuorejar SuLGnuapr op Osye st Yor “LNAGNAdAC Pue JAAOTAN uwoomiag addi diysuoneyar N:1 v YO"SINIGNATG 8°73 wr aydurexo 04 ‘A ua oan pares axe uo stoyp Jo Somnqunre Koy aavy iOu op rerp sodKy Aug yes Apug xeon SZ We SURG waUiaSeUEW aseqETeC quounmedap a4p 01 parejas are iysuonepax sadeueuy axp ut 99Ko}duso Or SHON “A ‘ounsu luo yspxo ue Aynua aaojduuo ue 94) p savers ayjod Auedwioo w 31 3g diysuoneya: You ~ SRIOM & ‘suounmedap v 10) 30M sn s9Xojduto 19 ‘adi drysuoneyas om via A jou 0} parejas Buroq s1_ wo spuodap Aanuo ue jo souaystxo ayp OYTO Saytoods (1: Wi) 2u0 07 Auoyy (ay) O} © SIePOW eG ore22 Meaning Entity ‘Weak entity Relationship Identifying rela Attribute Key attribute lued ‘Composite attribute © oS —<> —> -—=> « Derived attribute E R ey Total participation of E, in R 1 N Ei R ©] Cardinality ratio L:N for E,-E, in R (min, Fr Fig. 2.7 Summary of ER diagrameotation Data Models sbase Management Systems 243 (on) Conzaine PROIECT ContR01s >> DeraRroteNT an GoRKs FOR WORKS ON 0.1) |] epee DEPENDENT EMPLOYEE structural constriants onpuiayos auypary uo of woiBoip YF 6° Bt Orr A was || Lown wes s5v ALNdGNadda LL “LWaANadaG Iq “LNAGNAdaG AT TAKE aaa < )SLNAGNAdIG 9 auuua¥ep ‘NO-SHWOA\ °S adns w sey saXoydura rane sn atp aaye ‘Jonsed aq 01 + 51 LNAWLAVdaC Jo re svas0y “LOafOwd Pu LNIWLAVaSC WH € % unoyps ou ww 205 wresBeIP YS “T Pue JEAOTAWA woomog od EL @ ‘SDVNVIN sojdurexo uresep UA :sodAi diysuonjan Sutmorjoy ayy Asioads 2% g°Z “Bed Uy uMoYS se “o|duexD Jno UL siz ‘SUITES WoURBEUE aSOqEIEC ‘SIePOW Be ve2.16 Gut & 2. ER diagram for an Bank schema “BARC >: 'BANKBRANCH Ge Go) Ca) aie) | Bee. 08> ‘ Acar ([agsoute}-Cipe)) Data Models Ac => (> ~ Phone) [CUSTOMER a ‘aa Fig, 2.10 ER diagram for an Bank schema 2.6 INTRODUCTION TO RELATIONAL DATABASES 2.6.1. An Informal look at the relational model The relational systems are based on a formal foundation, or theory, called the relational model of data, The relational model is often described as having the following, three aspects: ‘+ Structural aspect : The data in the database is perceived by the user as tables, and. nothing but tables. ‘+ Integrity aspect: Those tables satisfy certain integrity constraints, to be discussed in the later section. Database Management Systems 217 ‘Manipulative aspect: The operators ava ihe user for manipulating those are operators that derive tables mportant ones are select, project iple relational database, the departments and employees database, is shown in ject, and join operations against the ions of those operations. . from a table, ‘¢ The project operation extracts specified columns from a table. +The oin operation combines two tab common column. ‘one on the basis of common values in Examples in Fig. 2.12, the only one that seems to need any further explanation is to have a common column, which table DEPARTMENT and EMPLOYEE do (they both have a column called Dno, and so they can be joined on the basis of common values in that column.) no Di. | Research pe | Admi ps | Sales Chennai EMPLOYEE Eno | Ename | Dao | Salary El pi | 5000 £2 | Dharshan | D1 | 6000 ES | Pradeep 8000 e4 | Praveen 000 Fig, 2.11 The department and employee database (sample values)(sajduxa) wof pun yoafoud was 212 Sty “suas fe Jo ansuaroesay Hurysin@upstp zofeu w st aigedeo Buyssoooud ros sty ‘sm Senay 0002 waned | ya za “B-t8-pu0994 10 -MOI at 1444 axe suraisXs yeuonejaruou wt suonesodo ay} “IsemTuOD coor |: despey | ca ta ‘Ag ‘swos snoy jo 3}qe aynsas v suunyan pus ‘Kfoarizadsax SM nog pure aaxtp 40 S2}qE) Om uo sareiado ZZ Sty ur uno! xp ‘ajdurexa 104 ‘smos J SIas UTEIOD Sayqen put ‘S¥Or aU 009 | unysieua | zg la asf you 's9jqet 2joys ane synsax pu spuerodo ays Te are suoneiodo aqp voip st sarensnyt Apedqo ose Z1°z “Bla wep aMtod roMPOUTY 00s wa | ig la foes | auoug | oug oud ‘suonaynsar om Jo mole ‘uiofe yo uonsafosd v ayer o1‘ajdusexa 305 ‘ayqussod S43 STU, - -qnoue oF indut auiooaq uo uonRrado au0 toss indiio ax — sage 4 ‘agp se 109{go Jo pury aures axp st uoresado Aue yo indino axp osm2099 au toxo sq,0 Tana poe LNAWIEAVa of awn | ea rewoy | za reuwoy | a ap BHT STON “eH nwa sures ay axey Kaqp astiedag saa ese wef | 1a | weoeya | wemasy | 1a 000 a | 1a = ‘Kepes | ouwug | ovg | soa | oueua | ova ou ong zan0 aerial sraofoug os | ia | wha | ta ‘oes | ong | ung | oug yeuuou, a reuway9 ta weuwan [aoweey | 1a 201d usu soa | sueug | oud mol TAAO TANG PUP LNA RLVdEIC 2p “9[durexo 04 “onyea ou voUtED w ane qPuURY1D,, = 901¢] 30H" LNAI {uo pue jt (ajqer soz axp jo more proMt 01) AAO TAWA AI wa INSINTAVAIC 219 WOH, Nor UoAID v‘oUoads 99 OL 9198 ‘suioiskg wwoURBeUW o5eqe°C ‘S19POW BEG are2.20 Data Models Database Management Systems 221 1 fora moment. There are a couple of additional points to 1e sample database of that figure: REVIEW QUESTIONS Tet us return to Fi ‘be made in connection wit 4s First; note that relational systems require only that the database be perceived by the PART=A user as tables. Tables are the logical structure in a relational system, not the physical fact, the system is free to store the data any way indexing, hashing, pointer chains, compression, itcan map that stored representation to tables at the 1. Explain briefly about E-R data model. [Nov’ 2003] Compare network mode! wi hierarchical model and so on—provided only’ ces [May 2005] fogical evel. Another way of saying the same thing is that tables represent an abstraction of the way the data is physically stored—an abstraction in which fmumerous storage-level details (such as stored record placement, stored record 2. 3. What does data model mean’ 4 sequence, stored data value representations, stored record prefixes, stored access 5, Name the different types of attributes that occur in the ER model : 6 1. 8 9. [Nov' 2006] Differentiate a weak entity set from a strong entity set, (May 2007 & Nov’ 2004] ; < cor pe [May 2003] structures such as indexes, and so forth) are all hidden from the 5. What is a derived attribute? Give an example, [Nov’ 2004] . ational databases abide by a very nice principle, called The information 7 presented inone ad /, Summarize various data models and highlight its feature, [Nov’2005] . Summarize the ER model notations. [May 2006] Give example for multi-valued attribute and composite attribute, [Nov’ 2006] table to another. In Fig, 2.11, for example, there is a connection between the DI tow of table DEPARTMENT and the El row of table EMPLOYEE, because ‘employee El works in department st connection is represented, not by & [Nov’ 2007 & May 2008] pointer, but by the appearance of the valu ; EMPLOYEE row for ED) ee you mean by weak entity set? Explain its purpose with an example. 1, What are the data models and how are they grouped? (6] [May 2003 & May 2008] So much for the structural and manipulative as ‘now we tum to the integrity aspeet. Consider the departments-and-employees database 2. Explain in detail any two data models with sample databases. Fig. 2.11 once again. [May 2003 & May 2008] 1. Each row in table DEPARTMENT must include a unique Dno value; likewise, ‘each row in table EMPLOYEE must include a unique Eno value. We say, loosel 3, How to represent strong entity sets using table? Explain with example. Jle DEPARTMENT and Eno in table EMPLOYEE are t ir respective tables. ‘that columns Dno 1 primary keys for t [Nov'2003] Bx jcal structure ofa database with the 7 m ble EMPLOYEE must exist as a Dno value in table aaa SerINoraeed i DEPARTMEN y employee must be assigned to an | } [Nov'2004] { existing department. We sa no in table EMPLOYEE is @ Discuss in det foreign key, referencing the primary key of table DEPARTMENT. respect to an ER enterprise schema. Construct an ER diagram for bank | 2. Each Dno value vo of the most importar es of mapping ct h [Nov'2004] ansactions, [8] [May 2005 & Nov'2005] ‘What is meant by re relationship types. ve relationship type? Give some examples of recursive [Nov'2005}‘eseqeted shed - pur - sayddng oy, suonoesues,, ‘smal, pue Seige] 2528 ‘seseqeie IeuOnE|ay Jo SINIONAS TAGOJA] TWNOLLV TAY OL NOLLOAGOULNT ¢- uaLav ‘TaGOJA] TWNOLLY 19 7- uN qqgqgdq [e00z keyul fg] “ursiexp yg anoge a1p 20) soIqun aepdoxdée vrssI0. (1) -swuaptooe popsoess Jo soqu poroyoosse sey 109 ype “ino a10W / 200 SUMO WOH Jo Yor9 sey rip Auedusoo aouesnsul 329 v 104 wresBeIp Yq Ue JoRNSUO (1) 91 [9002.40NI [8] -uesdeap yg ue Suudojorap a]ym paxspisuos aq 01 Sansst amp STENSNI St [9002.40N1 ‘jepow erep Jo souoBawo upeus amp {900z.40N] Is} skoy pure soinqimy ‘sis Aanug ‘sadA] “pug, suusoy SuMMoWOH 249 SUH}9C CT [s00z.40n > L007 Aeqa] [8] vone|su Aug w20y‘soyqen ony ye 32 plnom xo} “sangua om Uaaatiag 1x9 eur Eup suoNo1 Jo sada ayqissod ox {002 Sew Io] agua ye2Q (P) Wi ‘angune Koy pue sanyea ION (2) i ‘aangume paauiog (4) i aingume panqea-pinyy (®) 10 tp inoge ssnosy [cooz.4onl {sl “uonejau Auedusos”saXojduuo uonyB onoge ou Joy wesBemp Ya aM MEICL"OL cooz.s0nl ts -pjdumexa py japour euonwyas Soryou “ooryreaTy usswuDG aTENUAIOBI 6 1,002.40N * 9007 ew] [8] “paronpuos IEUTWUEXD pure SISO SNOUTEA atp Jo Ho] ¥ PUN ItoTEed Yow areID0ssY “s:O0190P oypaut Jo 195 v pur siuoqed jo 195 yptm jeidsoy v 405 wwesBeIp Ya We IONAISUOD SI9POW BEG weCHAPTER Se INTRODUCTION TO RELATIONAL MODEL 3.4 STRUCTURE OF RELATIONAL DATABASES ‘The relational model represents database as a collection of relations. ofa collection of tables, each of which is assigned a specifying a domain is to specify a data type from which the data values. forming the domain are drawn. 3.4.1 Basic Structure Relation name Attributes EMPLOYEE Empname | Salary ic Dao Praja 40000 tol ‘Anand 35000 | Projet Leader or ‘Tuples Kumar 30000 | Team Leader 103 Pradeep 25000 | Programmer los Praveen 25000 | Programmer tos Arelation schema R, denoted by R and alist of attributes: domain D in the rel Arelation (orrel by 1(R), isa set of n — tuples r= { he name of a role played by some n ofA, and is denoted by dom (A,)oF oy oOMPER “UROWUNOG pub pang cory gH PIO. junowe a i Nvor S0EV =n sic us wey ‘ean ae oor 102v ener Torv ean ay eH amen | SuEMTSUORTT wouIsoasa eH pmo our paHON ing aa e601 WON 3pauBOH a wApOO nono co waeOHR won 8 “app aarpeE woNvea oe TH pune va PoonpaRT fa wore a snus 06 vane a} Tipuanad Cj Twnonwoa mong eure qo0e9 ANnODD¥ pauaysopun aso soy A1ouad ayy ‘ouayos asoqorep youoymjax YNVE 24) sof wosBoIp ouayas. z°¢ Bt Tq uITROH SmeaIBwOIsNS waMomo" aqumunos | weeewon ousoss0 co) TT nvor wawousno Toe epee a woNvaal Sar [aoe | STUN aNnooo¥ ewsyos eseqeed Z'Ve en a a Le cet ne] POW [PUOH|NY OF WOHNORT vtay tabase Management Systems 35 34 Introduction to Relational Model pornower | customer-name | _Lonn-number ‘dans Lis cury L3 Hes Lis seston La ooes La Sou Lu Sait La Wits Lain ‘CUSTOMER | “customer-name | customer-street [aso Spring [rcs Senator [ou North [oie Send Hill [ Green ‘Walnut Main ‘Alma Palo Ao Main Varios Ped We Sami Preston Fig.3.3 One possible relational database state for to the BANK relation database schema 3.1.3 The Catalog objects that are of interest tothe sfstem itself. Examples of such objects are relvars, indexes, {sers, integrity constraints, security constraints, and so on. Descriptor information is essential ifthe system is to do its job properly. For example, the optimizer uses catalog information ide how to implement user requests. yrmation about users and security c ‘Now, one of the nice features of relational systems is that, in such a systerh, the catalog itself consists of relvars (more prec n relvars, so called to distinguish them from ordinary user ones). As a result, users can interrogate the catalog in exactly the satne way they interrogate their own data. For example, the catalog in an SQL system include two system relvars called TABLE and COLUMN, the purpose of which is” to describe the tables (or relvars) in the database and the columns in those tables. For the ’ 3aAO UN € | aNannavaaa ‘uNnoomoy | INAS NWNI09 | aWWN"aTAVE | atavE TPPOW [euorTe|ay oF woHIaMpasyy oe38 Introduction to Relational Model 3.3 TRANSACTIONS "typically involving several database sds to be able to inform the system when distinct operations are part of the same transaction, and the BEGIN TRANSACTION, COMMIT, and ROLLBACK operations are provided for this purpose. Basically, a transaction begins when a BEGIN TRANSACTION operation is executed, and terminates when a corresponding COMMIT or ROLLBACK operation is executed. For example (pseudocode): BEGIN TRANSACTION ; /* move $$$ from account A to account B */ UPDATE account A ; P* withdrawal */ UPDATE account B ; 1 deposit */ IF everything worked fine ‘THEN COMMIT ; 1 normal end */ /* abnormal end */ ELSE ROLLBACK ; they are guaranteed (from a ‘ornot to execute atall, even ‘even ifthe system subseque ‘transactions that makes the data in the database persistent.) 3. Transactions are also guaranteed to be isolated from one anoths database updates made by a given transaction TI are not made ly executes COMMIT, COMMIT causes, (0 become to other transactions; ‘and are guaranteed never to be canceled. If instead, all database updates made by the s latter case, the effect is as if the such updates are said to be commi the transaction executes ROLLBAC ‘transaction are canceled (rolled back). In transaction never ran in the first place. varanteed those The interleaved execution of a set of concurrent transactions is usually to be serializable, in the sense that it produces the same result as executi same transactions one at atime in some unspecified serial order. ase Management Systems 39 4 THE SUPPLIER - AND - PARTS DATABASE The purpose o (0 explain that database, in order to serve as a point of reference for later 5 shows a set of sample data values; subsequent examples will actually assume these specific values, where it makes any difference. ‘The database is meant to be understood as follows ‘© Relvar SUPPLIERS represents suppliers (more accurately, suppliers under contract) supplier; a supplier name (SUPPLIER_NAME), not necessarily unique (though the SUPPLIER_NAME values do happen to be unique in Fig. 3.5); a rati located in exactly one city. ‘| Relvar PARTS represents parts (more accurately, kinds of parts). Bach kind of part ‘has apart number (PART_NUMBER), which is unique; a part name (PART.NAME); color (COLOR); a weight (WEIGHT); and a location where parts of that kind are stored (CITY). We assume where it makes any difference that part weights are given in pounds. We also assume that each kind of part comes in exactly one color and is stored in a warehouse in exactly one city. sureun { sureuiex [ surruier | status ] cry | smpwewr| sureuer. | parrs_ | quantiry vumper | NAME Numper | NUMER st ‘Sia 20 Bony si PL x0 2 Una 0 st n a 3 Nein 0 SI B oo rt Pia 20 | Bunty sl 20 s Aad 30 | Dai si Ps vo SI 6 ‘0 2 Pl 300 rants [rant [rant_ | couor | were [ cry |] S fn ia NuMBER | NAME $s a my 7 Nit Red 0 st BR 200 m Bit Gea | ito 4 4 0 | a seen | Bhe no st i 0 | Pe saew | Red rr 8 oo Blu ao 6 og ed 130 Fig, 3.5 The suppliers and parts database (sample values)e002 90nd [8] -ojduexo ojqeuns © fay sndns pur Soy wSto1oy Coy Anoutad sur aip Buoure woNUNSIP yp URED. “1 a-Lavd ‘uonoesues], suY2q “OT ‘aata pue 219@1 256 UEC “6 {BopwwD Aq rea si ey “8 Igo0z Sew] “Koy ayeprpues pur Kay Azeunad ussMieg YSMBUNSI “L {9007 son) ‘Koy sadns pu £2x ayoprpuea ‘Kay CrewLid woaMioq UONOUNSIP 2) 2A1D “9 coo sevl ‘Koxsodns w pure Cay v uadisiag AoUasagpp 2xp St EULA °§ {s007.40N) “Kay uBtax0y pue Aay Arourtsd uaamioq 2ouasayip 2% + {s00z 40N1 -aydm wus ayn auyja “€ {ro0z27a) -skoy jo sadk quasagip aup AIA jenuassa Aye st MUM “TZ {2007 son} -ways(s quouraSeuCU aseqriep Ut pasn say jo sad wuasaytp ayn ‘aydurexo afqenins yim Ayroads “1 v-navd SNOUS3ND M3IAZY SNOB MBIA We TDIeNG wauRouyy S50qe1e0, “yd syynoqued e pue sayddns zojiopzed w woowog daysuoneyas se popuvar ag uo yuowdrys © pure ‘se poprnau aq uvo sured pur s1oxjddng te ve suauadiys ou wun ‘sg 241ddns ‘sou ddns auo sapnyour ¢'¢ “B14 Jo 2ssqp EN SINGH ibd psuna sw J0 80 tw st aioyy oumsse 2M “(AILINVO) joquint wed © “(USEWALN: WaTTaANS) “‘Kypureu) SHAT Idd SUNWAIHS 11 mon sut oxy ‘ejdurexa. ‘woy aaqidins oyssods w sy Surypads Sqeot80y souyaBonsreaja Om J2mHO 94 UH] 01 aguas w UF SaAIas y “sIuOLUdrYS syuosaudar SLNAWIHS TUAPA ¢ TPPOW feuoNEI2y OF uonaNpONU oreCHapTer - 4 Types Introduction Values Vs Variables Type Gener SQl Facilities, BBBB‘stoquinu 2oyjddns Suysedwuod 404 “Uo os pun PInow aM AGWAN YAN Tdd NS addi pauyyep-ta =» ssorerado ouyjap Aiqeqoxd yo asvo ayy ut senu0d Kg, (pauioddns 10u are siafayuy tuo suonerado Suns ‘sprom so4po ut) siafanus uo suoneiado uyns Suywuoprod 40y sioyesado apiaoad tou sop] EY Fuyoy20d 109 “uo os pure jorsado saprnosd ogye| ‘Buzeduoo 103 ‘uo os pu ‘.>..‘,=,, storerado sopraoud marsss ay, « |YADALLNI 244 pautyap-wrsés axp jo aso oxp ut “apdurexo 30g. “(eda emp 30.99 01 pasejoap st veyp sorourezed w sey uonsanb wr soqesado orp EIN ‘fyostoasd “uvaw am 04a yeyy 07 paurjap, & axauy) adKi yey 209 pautyap srorwsado aup Jo suvau a Atajos uodn parezado 99 we9 ad uoatd e Jos yp ‘uonsanb ut ads aap Jo sanqea 01 pordds oq Aipryen to wey) suonesado Jo 19s pareioosse ue sey adf) and UY ‘amnquine jeuonyar Sutzej20p 403 s1seq 2p sw past 2q wo ‘pauyjap-iasn 40 -tiarsis st 1 1oypayn Jo ssoypmefas ‘roxoosteq a6 ky “pauyop-128N St YAGWAN Wal Tdd Ns Pu peuyap-woisés axe YVHD pue UTOALNITEN auunsse aj “pauiyap-19sn 10 (1 rss JoUa st adh uaa? Ay ‘syxonuod feuonejar uy Swouoxey 01 Ajuo inq ‘pauoddns st Bud Huong.“ ~pamnaaxe aie uD 19 ©} siuowUsisse jenplAypUl ay) Jo = LAS soo] auayeis ward “sad osamp $0 12 303 2 wotodeid possoooe 2g 0} wioyp stole (SBuny JOY BuowE) reHP yeuuzopu ‘or panajas ‘osty “(sioqum (i) €002" 1s. uy urede paddoxp 29 [11M pur 2661 “JOS UE Pappe 4 ONIKAVA Le Pur Lid “t ch says wOUETETE 989 odds osye 227—1.9¢fO IOUVT RAVNIG 20) COTE ‘LONGO IONVT YALVAVHD 26} GOTO AALOVAVHD 44 VHD ‘aduuexa 0} .qqe ‘snnyap yo oquinu Wy @ gvorld ()L9ardO IOUT WALOVAVHD WAMGINI INITIVNS ONIAMWA ] YALOVEVHO aNVISAINIL YADAIN (1) LOarO AAV] AUVNI aL (b4) INIA (a) LONIAUVA ] Lid alvd ‘d) OPYAWAN Nva1008 ssodAi uraqing Kroreunydxo-y[98 $59] 20 lop aup soprsoud TOS soddy wrung saumovs Tos yt soplo asranad Ur au0 wDAI3 3m 9 d4oUDR B 2 {sus J9m0} PURO sON0} 2 sowwsouad adfi VARA atp Jo 9509 a4p uy ‘aydurexe J04 “(uonsend ‘uy Joqesouod afi 2xp Jo uoneoosur wa pauEago adKs oyoods krona on éydde ys woRSoND io) 1 pla paverposse sxoresado yorouod addy Uaat KUy :9I0N, (1 ‘S “1 ‘sz'ze Ee ‘LZ “SI “6 “6 “S °C) AAOALNI AVIV = SATVS: ‘uosuduoo ssyyenba plyea e st 2154 PUY “G “61 '6z°ZE “EE ‘LT “ST “6 “6 “S “Z) UAOLNI AVY — SITS ‘yuOUUsse prea sy asoy ‘a]dumexo 40g “Xjddu os stoyerado ‘jqeuea opnasd v se pasn oq osye aysttt i 1 9q 04 stiaddey worn anyea ssoo0e 0} pasn aq 1ydtu uoyssouda atp “2staoH7T AvAnAV-AadG Jo anpwa senor md v Ajoads 01 PAS ‘sok,ij 46, ‘Types 3 For character string types in particular—CHAR(a), CHAR VARYING(a), and CLOB(w)—the type checking rules are quite complex. Full details are beyond the scope of this book, but we do need to elaborate briefly on the case of fixed-length character strings (i.e., type CHAR(M) in particular: .R(n2) are compared, the shorter the same length as the strings *P2" (of length ‘values of type CHAR(al and is conceptually padded at the right with blanks to make longer before the comparison is done, Thus, for examp! two) and ‘P2” (of length 3) are considered to “compare equi length n2, DISTINCT Types SQL supports two kinds of user-defined types, DISTINCT types and structured types, both of which are defined by means of the CREATE TYPE statement. Consider DISTINCT types structured types in the next (we set “DISTINCT” in uppercase to stress the poi ford is not being used in this context in its usual natural language sense). The following is an SQL definition for the DISTINCT type WEIGHT: CREATE TYPE WE! AS DECIMAL (5,1) FINAL; implest form (ie., ignoring a variety of optional specifications), the syntax CREATE TYPE AS FINAL 3 all, One consequence of this omi ‘ype—or a structured type, com 3. There is nothing analogous to the Tut case of type WEIGHT, for example, there is no way to specify that for each WEIGHT value, the corresponding DECIMAL(S,1) value must be greater than zero (!) ot less than 5,000. base Management Systems aa “The comparison operators that apply to the DISTINCT type being defined are precisely those that apply to the underlying physical representation. Note: Apart from assignment (see point 8). other operators that apply to the physical representation do not apply to the DISTINCT (ype. For example, none of the following expressions is valid, even if WT is of type WEIGHT: WT 147 WT 42+ WT “Selector” and “THE.” operators are supp variable of type DECIMALS. corresponding weight value; and if WT is a column of type WEIGHT, then the expression DECIMAL(WT) returns the corresponding DECIMAL(5,1) value. Hence, the following are valid SQL st DELETE FROM PARTS WHERE WEIGHT = WEIGH EXEC SQL DELETE FROM PARTS WHERE WEIGHT = WEIGHT (:NW); EXEC SQL DECLARE Z CURSOR FOR SELECT DECIMAL ( WEIGHT FROM PARTS WHERE WEIGHT> WEIGHT (NW); 7. Withoneimportan types. Note in particul and values ofthe unde 4.7): s DWT DELETE FROM PARTS WHERE WEIGHT =14.7 EXEC SQL DELETE FROM PARTS inval EXEC SQL DECLARE Z CURSOR SELECT DECIMAL ( WEIGI FROM PARTS WHERE WEI )AS DWT [> NW; /* warning — invalid€ Je S71 J0emaNE NIDE 4/ Ix Joeinqune x 'Z=XNIOSE'ST Las Je STs amqune NIDA «/ Je JOaInquae X .S9A2590,,4/ !XNIOGA'ST =X 1a Ie dyoamnqun x , 'Z=xalas jd JoomnguM x ,S9019590,, 4) tXd=Z1as pipes axe noadsau LVOTA PtP “LNIOd DASANIT S244 Jo ‘a|duexa 104 ‘stoyeiodo 4H S.C [POI] Jo Tey 0} snodoyeUE Autjeuonsuny oprvoid yey ‘ioyynuu U0 pur JaA12sqo 40 siojesedo paeroosse omy Jo uontuyap onewioyne sasnea uoN z 119% se uonejuasasdas zojod v asodxa on, gy 219}, + agAL dowd :adAi pauyap-iasn w dosp 0} pasn st rowayeys Sutmvo|foy 24 (oe Url) Lmaav CIM ‘LM) Lwady (Cia ‘Lr Lavaay (cer LM) Lwady 1e9 suoyeiado JeuCMIPPY “OL ‘sod LONLLSIG ‘Uaamjaq 40 “wos ‘0} Sups9auios 10} pouyap aq ose uvo sxoqeIado ASV. ‘suonwiado Vdd Pue LAASNI 0} édde sysewias snoBopeuy )AAGWON LYVd = AIGWAN LaVd FUaHA WOU sLaivd AN LHOTEAN L osye st Suyaoyjoy ayp ‘oAaMoH] (ana90 |}}m. woyas909 oy '(.1d.) YIEWAN'LAVd = WEAN Vd FAH. SLIVd WOU AN: OLNI LMG SV (HOM )IWWIoga Loa7as a64.10 4 1. Operators in addi ‘Types “The comparison operators that apply to the structured type being defined are specified by means of a separate CREATE ORDERING statement. Here are two examples: ‘CREATE ORDERING FOR POINT EQUALS ONLY BY STATE ; ‘CREATE ORDERING FOR LINESEG EQUALS ONLY BY STATE ; EQUALS ONLY means that “=” and “#” (or “<>”, rather, this later being the ‘SQL syntax for “not equals”) are the only valid comparison operators for values of| the type in question. BY STATE means that two values of the type in question are ‘equal if and only if, forall, their ith attributes are equal. ‘No selectors are provided automatically, but their effect can be achieved as follows. First, SQL does automatically provide what i calls constructor functions, but such functions return the same value on every invocation—namely, that value of the type in question whose atributes all have the applicable default value. Forexample, the constructor function invocation POINT () returns the point with default X and Y values. Now, however, we can immediately POINT () X (5.0). Y(2.5) Here is a more complex example: LINESEG () . BEGIN (POINT() .X (5.0) . ¥ 2.5) END (POINT() -X (73) . ¥ @8)) ‘Note: Constructor function invocations can optionally be preceded by the noiseword NEW without changing the semantics. For example: NEW LINESEG (). BEGIN (NEW POINT() . X (5.0) . Y@.5)) END (NEWPOINT () . X(7.3) . ¥(0.8)) 5. Strong typing does apply to structured types. 1 to those already mentioned can be defined (and subsequent dropped) as required. Database Management Systems : 8, Structured types and orderings can be dropped. Such types can be “altered,” t00, via an ALTER TYPE statement - for example, new attributes can be added or existing ones dropped (in other words, the representation can be changed). Type Generators ‘SQL supports three type generators (the SQL term is type constructors): REF, ROW, and ARRAY.” Here is an example illustrating the use of ROW: CREATE TABLE CUST (CUSTNO CHARG), ADDR ROW (STREET CHAR(S0), CITY CHARQS5), STATE CHAR ZIP CHAR(S)) PRIMARY KEY (CUSTNO )); STREET, CITY, STATE, and ZIP here are the fields ofthe generated row type. In general, such fields can be of any type, including other row types. Ficld-level references ‘make use of dot qualification, as in the following example (the syntax is ., where the must be row-valued): SELECT CX.CUSTNO FROM CUSTAS CX WHERE CX. ADDR. STATE =*CA’; Note: CX here isa correlation name, here we simply note that SQL requires exp correlation names to be used in field references, in order to avoid a certain synt ambiguity that might otherwise occur, Here now is an INSERT example: INSERT INTO CUST (CUSTNO, ADDR ) VALUES ( ‘666°, ROW (1600 Pennsylvania Ave.", ‘Washington’, ‘DC’, *20500" ) ); Note the row literal in this example (actual formally, there iso such thing as a row li is a row value constructor). that should be “row literal,” in quotes— in SQL, and the expression in the example“70s tt soda pasmonns pur sod&y ounsp mnoge ured “yerep uy storesous8 adi oxy uroyxg THeIp Ur a1qeHEA sa angen oMp UTE|E, a-uvd esonyen Jo 198.69 yuwau st yyy {TOs 4a poutoddins sad wrajng axp Jo aunos uonuayy {Suid Suons yo asn amp woRUIA (aowerauad ad £q yoeou st yey “arqenr, pu anyea aujoq y-nivd SNOLLSIND MAIAaY ery ‘swaps wowntiouey aseqeIeC soxpta onsonb un ad AY AAW 40 A\OW aH 404 pauyp row 9580 yoIYN tt pauyap YoU st wostredtOD IY 40y a4 UatHD|O te SOnqonu uonsanb ur adh AVeRIW 40 MOY aH Ss9puM—sodd AVAPAY ue AO Hog 10} ‘Kydde op stonesado uostieduon Aiyjenba pu uounudisse req Busou Kq uonsas sp Asoo 344 $.¥9902Z.= ONAL TUTHA, ol=(Isatvs das savs WALI divaan ‘([ov'o'o'''00'0'0'0'0] AVY OZEPXs ) SANTVA (sas ‘ONWALL) STTVS” WALI OLNI DEESNI Sol <{€] S1TVS WAHL STIVS WALL WOUd ONWALLI LO31S -ydurexo jojo = toured, TOE] exe, 30) [2A sojdexo wos are MoU 3194] uoo9s ay ur (doronnst09 anjea Seas Ue oaie ayy ON “3I9R SITVS WAL AN ‘uanpea oup wani9H (8) A LTTWNIGVO worssaudxa ou [Jose paouougjau 2q uo pur sfostoaud axe squaUIa]> asoqt uOKN “(0 < w) suoutaja u Koexs streiuo9 » Jf “punog saddn payyoods ayy weqy soreax8 You Inq o1dn (9 anjo {ya a wo afqeea uonefar an on pousisse st —> ayn Jo WSABOyRSAL AML < ‘suogesedo e1qo6| “suorssaudxo yuonbasqn: Fy > | dwar ses + som rode ‘3]qeuwA oHjau Kresodwioy ¥ 0} opewt aq sKeIATe ISM WOUUBISSY — < nea asaya torssaudxa we Kq pamozfog ‘uwsfoud jenuanbas ese aonb amy < siuoumuBisse Jo souos v Jo Suns -soyzanb resedo MOUSSE ONL < (>) wayeisdg suamudissy ( ‘xapduroo ssoudxo 01 Kem wworuonuoo v saprncud (>) Be sarwwwe 9 ’ £ 1 1 1 £ z 1 a <|sa < olalae Benue Aion0 Query Langage Database Management Systems sa > Uses Null Values 53 MODIFICATION OF THE DATABASE (@) Nal signifies thatthe value is unknown or doesnot exit. eat D (b) All comparigons involving null are false by definition. In relational algebra, a deletion is expressed by xamples aston rer-E ee tes, T-170 Downtown | 3000 Jones risarelation and | vat fee | ine ue Eis a relational algebra query . Example nner Join ([><]) Delete all of Smith’s account records. Loss [><] Borrower Toan-ao] branch-name | amount] customer-name 1-170} Download | 3000 ones L230 | Redwood | 4000 Smith Left outer join ( Loan }X{ Borrower Toamae [ branch-name [ amount | castonername Te17@ | Downiown | 3000 [Jones Lm | Redwood | 4000} Smith v2 | Peirdge | 1700 | Null Right outer join (X) oan [XL borrower Toaw-s0 | branchcname | amount | customer-name L-17 | Downtown | 3000 Tones. L-z0 | Redwood | 4000 Smith Lao | null ull Hayes Fall outer join (XC) Loan PE borrower . Toas-no | branch-name [amount | customeraame] ~ 170 | Download| 3000 Tones Lm | — Redwood | 4000 Smith L-2m0 | Perrridge | 1700 Nall Lass: | nuit ull Hayes depo: 5.3.2 Insertion 1 & depositor - 6, (depositor ) To insert data into ar result isa set of tuples to be le to be inserted or write a query whose Insertion is expressed by ReruE Where, risa relation and Eis a relational-algebra expression. Example To insert a fact that Smith has $1200 in account A - 973 at the Perryridge branch. account «account U {(A-973, “Perryridge”, 1200)} depositor < depositor U {(* A973)} 5.3.3 Updation Tochangea value in a tuple without changing all values in the tuples, the generalized projection operator is used. ) whiere each F, is either the i attribute of, ithe i attribute is not updated, or, ifthe attribute isto be updated. F, i an expression, involving only constants and the atributes of ', which gives the new value for the attribute. rely,4,10 'y ut punog 30 2ay5 at uo Sumpuadap (4) rou pu ('g) 104 430) “€y pur '3) aansouu09 30 dn apour enuioy w uy punog 40 aayy st} ajqeurea ajdny eyo guano UY 2A 204 5 wore West Mp g eM w w 2|qeLze opdm vo soUDUMODO UY « “sajna BuysoI]9} dxp oF SuIpioode punog 40 aay se BIMUUIOy v UI B]qeHTEA 2jdm iN Pue TenuarsiEa ZL¥'9 ‘wore we Jo anpeA tpn payfeo st 1dr Jo uoneurquioa oytsads © 40} ASTVA YO ANAL J2yH2 oF svenPeAa swoye Surpodard amp Jo YOR, uso — 3 azoya a‘haoo 49.9. dO V" 1 uo amp Jo wow uy “¢ soresado jeuone|ay ~ gO soBues{ yoIy uo woneyay amp Jo sainqumae — g “soBuer 1 yorya Uo uorrejas a Jo sainqume — y auaym ‘> ‘=} uoneiado uostediuoa oip Jo 2u0 s1 40 ‘ureu asoy uonejar ayn se 4 ayqeuen 9} atqeume adm — urea wonejas— Cx mangg amp Jo wo Uy “T ‘Buywo|]oj xp Jo au0 aq wea swore smjnopa’ayeaypaid jo dn apew st wymuasoy ¥ luuoy ayp Jo snnoyeo jeuonejas ajdn yp Jo uoyssaudxo yesouad Vy ‘seinuuos pue suojsseidxg 49'S $$ Sento) pute sucyssauda 1 b9's ers ‘swaishs wowsTounyy aseqeed {000'0s < Azeyes3 pue (1) AAAOTANG / SINWN? ANYNA} “saynqume jo atuns anaunas oy, 89 {ooo'os MODE -yuoaied ¢ Kq posearour 9g 0} are soowey tre exp pue ‘ape Sule are squouiXed ysazonuy yoxp asoddns “uoresodo ayepdn 2y) Jo a5n ayp ‘arENSMITE OI, aydumex ayfo-3 “Woe (@'o-an oa ‘1 sajdiy yoryss sasooyo yeup wonypuos woHI9}28 p d 2104 suorssaxdxo Sumo 2 agn ta am “uray équo atepdn oy pue wio4 sof) auOs Popes oY TURAL IM HL sBenduey GanQuery Language 4 ound in a formula F’ of the form [Ail free ocourrence of tuple variable tin F are be F=¢ PCy ‘The tuple variable is bound to er specified in F" = IfF isa formula, then (3 1)(F) tuple variable ‘The formula (3 t) (F) is TRUE if the formula F evaluates for some (atleast one) tuple assigned t6 free occurrences of t otherwise (3 1) (F) is FALSE. + IfF isa formula, then (yt) (F) ple variable Bg, Retrieve the name and adéress of all employee who work for ‘Research’ department {tFname, tLname, t.Address/EMPLOYEE (\) and (3 ¢) (DEPARTMEMT (4) and enname = ‘Research’ and DNUMBER=tDNO)} Using the Universal Quantifier Eg. Find the names of employees who work on all the projects controlled by department number 5 {ecLname, e.Fname/EMPLOYEE (¢) and ((y X) (not (projeet(x)) or not (xDNUM=5) of (G w) (WORKS_ON(w) and w.Bid = e.Bid and X.PNUMBER = wPNO)))} Safe Expressions A safe expression in is guaranteed to yield a number of tuples as a result otherwise the expression is called unsafe. Eg. {not (EMPLOYEE (0)} Is unsafe because it yield all tuples in the universe that are not EMPLOYEE tuples, which are infinitely numerous. We can define safe expressions more precisely by introducing the concept of the domain of a tuple relation calculus expression. This is the set of all values that either Database Management Systems FI 15 ‘appear as constant values in the expression or exist in any tuple 5 jn the expression iy tuple of the relations reference An expression is said to be safe if all val the expression. sin its result are from the domain of 6.4.2 The Domain Relational Calculus ‘The domain calculus differs from the tu rs from the tuple calculus in the type of variable used i fonmulas rather than having variable ange over tples, the variable range over cic values from domain of attributes. mao An See ‘of the domain calculus is of the form {x, x, .- are domain variable that range over domain of attributes, x,/COND (5. COND ~ condition or formula of the domain relational calculus. A for main re cal lus. A formula is made 1. Anatom of the form R(x,.x, R—name of are This atom states that a list of values of < whose name is R. x, value of the i* attribute value ofthe tuple. 2. Anatom of the form x, OP x, x) must be a tuple in the relation OP is one of the comparison operators in the set (=, x,and x, are domain variables. 3. Anatom of the form x, OP C (or) cops, C-constant, SSR) Eg. Retrive the birthdate and address of the employee whose name is “ABC” {c,d / (Za) Ge) EMPLOYEE (a,b,c,d,¢) and a="ABC’} where a= name b=Eid ¢=Dob d= address e=dno or{2002 .son] [9] (OUD sun: Tevoneler any. pus (OG) snynofeo yeuoneas uewog uoowog aye (eooz en) [+] dooz.40n £002 X61 's00z.40N ‘9902.0 ) “sojdurexo ajqviins yim suonessdo ‘saBenBize] Axang 1n0q [soz sey) "aga8e feuonsjar ur suowezod Hof [pe urejdxg [¢00z.40] [3] “odurexa yeas “elds quoniedo wqa3yefeuoneas © w wopeiodo of 21n9, moped on or {so0z.s0Ni I] ye9p ur wonesado ronpaud usu ayn urdxo ‘ojdumxa am [eo0z ew] [g] -w1938,e pouonejas 91) ur Suoneiado peusuepemy atp ssnosiqq a .piva {coor sey) aSenduey Conb jo uoneoypsse ap ox oSonBuo] ‘Auang) auyacy [9002.40N +9 sooz sey] “waged ye ‘ur suorresado aya 35 {a0 pure yorasoy.=d Yee) / Ps} [9002 ey) “aged euonyjau ut pasn sroyezodo snotzea oun Eig {sooz dey) “suoneyay wiof snouea atp Isr [9002 de 2p ¢00z .0n) époung}ad st uopesado ,wror - jem, Mop] [roqumnudop-1 uoneoo4 adop+-b oureup-d] (007 Aon} ‘soldurexs ins zausuv ano apensnyt ceuge8}e uonear ur ‘uonesado ourouar sty, pur (ibd) LNAWLavaaa pue (orP''¢*) aaAOTaWa GD) @E V- Riva : ure on 949K _Hotwasay, 91 405 410M os soaKojduo [J Jo ssamppe puE SNOLLS3ND MaIATY neers om ((@9'q*,oav,) 73014 us SISAS TOUTE oR Benue KonBREEEBBBBBBBEBBBBBEEBBEBB Carter - 6 SQL Funpamentats Introduction Advantage of SQL Parts of SQL Domain Types in SQL Data Definition Language Data Manipulation Language Basic Structure of SQL Expression Rename Operation Tuple Variables String Operations Order by Clause Set Operations Aggregate Functions Null Values Nested Subqueries Complex Queries Views Transaction Control Language Joins Embedded SQL Dynamic SQL ‘Advanced SQL Features“suonems rardutoo oqpurey ueo uso} 0} sea pur ajduuns Suing aqyn aBendup) ayy ‘stop 2q Pinoys Pauinbas st ya sayytoads 1g asnedsq st sty, OU 10 sax. ‘u20q sey Surziuindo sompoyrs Jo ssoypmea ynsau aures ap ‘V Ajreusoruy porawsay dun st Kem axp Jo iwopuadapur st 9 “peSuoy 40 paperfdn 24 01 spoou waskg uowoBeuyyy'asequieg Bux (columnnamel datatype(size), ‘columnname2 datatype(size)....);“ssn wouy ayy uy wade yy suonejx tp Jo sainque Suxafonut oeoqpard vo ss}su0D 3, “eigad}e euonejas 2xp Jo ayeoxpoud uonsoj9s ay) 0} spuodsatioo asne9 AZoM a4, “uorssaudxo ayp Jo worienjeao ayp ut pauvas aq 0} suorreax ay sist “eae uoneja4 aun Jo wonaiado jonposd wersau axp o1 spuodsazio9 asmeyo wos LL ~Asanb v Jo 3ynsoa at ut pasap saynqunye ox "aqa8fe jeuone[au ap Jo uorreiedo uonsafoud ap 01 spuodsait09 asmeyo 199 raya pu wos o2jas : sasnvyo aamty Jo sisisuo9 worssadxo JOg ue Jo amon 21seq aH. NoIssaudXa TOs 40 SUNLONULS OIsva 9 001 =Pis ayaa eaXojduis woy a}9PCTOS agdueorey ‘ axayaa wroay ay9]9p sous Pueuuo9 212jop Bussn parojap 2q UD sMoy Puvmwo3 area SOMV, = atueud asaya QOL=pI9 198 aaKo|duss apepdncTOS ayduioxg suo “onjoapyoyy 398 ayepdn om aU souls “pueurtwos owpdn Sutsn £q apeus 9q uvo saBuey puraiuio> ayepdn {0001'0s*. GY, sonpeA saXojduro omy wasM asap sous aumppnays 9148) 94) MOEA 01, ‘s0Kojdurs apqm aywounar ayqes ayeouny sus “amyonans ayy ureias Inq sprooo4 2up Stapp 01 Pasn St STL pueummo> ayeounay, -aumjonus se jjam se swuatuod atp a}afep {jun puEUIAIOD STH :5010N, ssokorduraaqqus doap apqea doxp xods ‘]qea 8 ar9|9p 01 pasm'st puBuURHOD sq puruia0 dosg s((@asquinu 230) ppw 2aXojdwoaqes 49916< TOs ¢((Qasquinu pra)gppou s0X0jdur994qe A0HTe 2148) 22318 (ozis)adhvevep Zauwnnumyoo sou ‘(ezis)odhrmep [oureuuumyos sou) ppe 2146) 491TE sous ‘suonruyap uumyoo Sunsixa Kp{pout 10 UUuN,OD MOU PPE O} PaAsn SE If puemran0; 9915 select Employee.name, address from Employee, Department where Department.name = “Research” and \ Department. Dnumber = Employee. Dnumber; @ ‘SQL>select Exame, E.address from Employee as E, Department as D where D.Dnumber = E.Dnumber; To rename the attribute seleet ename as Employee_name from employee; Database Management Systems 61 69 TUPLE VARIABLES 7 ‘Tuple variables are defined in the from clause via the use of the as clause. SQL>select B.customer_name, B.loan_no, L.amount from borrower as B, loan as L where L. Joan_no = B. loan_no; 6.10 STRING OPERATIONS operator for comparisons on character strings. aracters, Example: Find the names of customers where the 1S 2 characters are ‘Ba’. SQL>select customer_name from customer where customer_name like ‘Bat’; Find the names of customer where the 284 character is ‘n’ of ‘a’. SQL>select customer_name from customer where Tike ‘_1%6' or‘ at’; . upper case characters do not match lowercase characters, and vice versa. SQL supports a variety of string operations such as (@ Concatenation us P orst (©) Converting the string into upper or lower case upper of lower () (©) Find string length (strlen ()), extracting substring, (substr ()), ete.‘oureuyoueag kq dnos3 unoo9e mous (sounjeq) Ba ‘suTEN Youresg IDapeSS (8a) nb y09 Aq parsayos Sisou ToMRSIP [fe SUM IT (sou awwonidnp sor ado wow U tuo ayeiado ydaoxe pur jas do wigaSye yeuorejas ayy oF yerodo TOS UL, SNOLWHAdO 13S z1'9 ‘29st plo ‘asap Aroyes kq sapu0 saKoiuro moss + PAPSseleet branch_name, avg (balance) from account ‘group by branch_name having avg (balance) > 2000; 14 NULL VALUES * sa about ‘% Null signifies an unknown value or that a value does not exis. the use of null values to indicate absence of information value of an attribute, ‘© The predicate is null can be used to check for null values. xampl Find all toan numbers which appear in the loan relation w ' for amount SQL>seleet foan_no from loan where amount a for example axample: 5 + null returns m ‘onsider the unknown (null) value used in boolean expressions as 1. OR- (unknown or true) = true, (unknown or false) = unknown (unknown or unknown) = unknown. 2. AND- (true and unknown) = unknown, (false and unknown) = false, (unknown and unknown) = unknown. 3. NOT-- (not unknown) = unknown, & — Forexample, find total ofall loan amounts. SQL>select sum (amount) from loan; Above statement ignores all null amounts. The result is null if there is no non- ‘null amount. All aggregate operations except count (*) ignore tuples with null values on the aggregated attributes. base Management Systems 6. AS NESTED SUBQUERIES - SQL provides a mechanism for the ne is a select-from-where expressi A common use of subquery set comparisons and set cardi: Set Membership SQL uses in and not in construets to set membership tests. of subqueries. A subquery in another query. s for set membership, IN i, The IN connective te produced by a select c SQL>select distinct customer-name from borrower ‘where customer-name not in (select customer name from dé Set Comparison Fad the nares ofall branches that have assets greater than those of atleast one SQL>select distinet T, branchname from branch as T, branch as $ where T. assets > S. assets Same query using > some clause. SQL>select branch-name from branch where assets > some (select assets from branch where branch-city = “Chennai”); and < > some in. The keyword synonym to some in SQL.soumew sums ‘tp ur sajdny ayeorjdnp Jo aouaysixa ay sa) 20y pasn st jonsysueo anbyun you * wot), = aE yourg-ruNODae ue ou ryNOode = OUsTUMORDE “y put sureu-rouoysno “y = ott foutoyseS “1, axaya 2 Sv so1sodap ‘tumos.e OH atueu-rouorsno y 139998) ambqun a9q|st .L 88 sousodap wroay sure rowrorsm “1 199798}8o} suns axp 905 01 SHosn [Te 309 21QBNISOP 30K (suoojas 20 wont) $919 UE WO, BIEP JO MIIA TeorBoy w Jasn oyp SAIS YeYR SMAIA 149 ‘onqea‘sourepeq”xeur = doueyeq Iun0I0e a20y sourejeq” xe unoooe toss ou Mode 199I9s unosoe wos (Gouryeq) xeut 309}08 ‘sv (anyea) souejeq” xeur yA ‘patsofos aye wioyp Jo je ‘soueyoq wumuuTxeUr aus ayn YIM siUNOGIE AuELE ase aNOK, ‘sounpeg wrmumpceus 4) YAK spunoade sio9jas YoIKAA ‘zanb Buymoyjoy aq sapIsto) “sio9o astiojo maya amp YoTyAN Ur Azanb a4 oF Uo 9 WoNIUYaP asoym *NaIA Aresoduioy B BuIUYap Jo dem w SapIAo.d IsMEFD IA SAL, ome WAT sou Pu aureu-tyoursq soinqiame atp yum Bav-youBsg pauree st ynsax Asonbgns ax9H ‘000z < aoueyeq-Baw aaaqsa (sourpeg-Bav ‘sumu-yourig) Fae-ysurag se (oweu-yoursg 4q dnoad yunooe wou (aotejeq) BAL. “Sureu-youeig y99[98) wosy douejeq-Bav ‘ueU-youRIG 199]98ereate view customer_details (cust_name, customer_no) asselect customer_name, cust ‘Selecting data fram a view SQL>seect * from customer_details; Updation of a view ‘Views cau also be used for data mani insert, update, and the delete operat The views on whi 1er_id from customer; e., the user can perform ° rion can be done are called that do not allow data manipulation are called ° ‘When you give a view name in the update, insert, or delete statement, the ‘modificaion to the data will be passed to the underlying (main) relation. pase Management Systems 6 F the view to be updatable, it should meet following criteria 1, The view must be created on a single ta 2.” Primary key column of the table should be included in the view. } The select statement used for creating a view should not include distinct, group by or having clause. 5. Select statement used for creating a view should not include subqueries, 6. It must not use constant, strings or value expression like total/6, Destroying a view ‘A view can be dropped by using the drop view command, ‘Syntax drop view view_name; Example ‘SQL>drop view customer_details; 6.18 TRANSACTION CONTROL LANGUAGE * The TCL statements give you flexibility to undo transactions or write transactions to the disk. ‘Transactions provide consistency in case of a system failure. Weare having 3 TCL statements 1, COMMIT ~ Ends the current transaction and wr ‘permanent tothe disk. 2. SAVE POINT n - Marks a point in the current transaction, 3. ROLL BACK [TO SAVEPOINT a] ~ Ends the current transaction by ‘undoing all changes if a TO SAVEPOINT Clause is not used. to the save point if the clause is used, removing the save point and any change after the save point. All changes We will see the Data control language (DCL, statements under te topic security in the next session.ee tany Oy 2u0 ‘Pounbar axe siuowres om, zou Ages stag ai ‘tah nou 188 tonb eo ynsax ayy ators ot ryasn ayo st ton xojdtooe Suman uy, “yunoace se Buses ues a4) Sey Yor TnoDot dura ojqe mat soya UoUaIS axoge ay yunos9e ayy wunodoe- Not null constraint > Check constraint (Not null constraint (cg) SQL>ereate table employee (cid number(S) constraint emp not null, ename varehar(2));, emp - constraint name (itis optional) is used to enforce thatthe particular column will not accept null values, Check constraint Itis used to specify the conditions that each row must satisfy (eg) SQL>ereate table employee salary number (5) constrai values that ® 10), eid number(5), sal check (salary > 500)); is optional) {fa value less than 500 is entered for salary, then it will produce error saying constraint has violated. Note: not null and cheek constraints are column level constraints. sal - constraint name tthe‘mnozon saouasafes (ou~200)Kay wSjauof “3]qea poouasayar 91p Jo soinque Koy sound oxp soouasagas Koy wia10j w ywjap ereate table customer (customer-name chart char (20), customer-city char(30), primary key(customer-name));, ‘SQL>ereate table depositor (customer-name char(10), account-no char(10), primary key(customer-name, account-no), foreign key(customer-name) references customer); ASSERTIONS © Anassertion is a predicate expressing a condition that we wish the database always to satisfy ‘© Anassertion in SQL takes the from SQL> ereate name > check < predicate>; © Whenan assertion is made, tests it for validity and t again on every update that may violate the assertion © Asserting forall X, P (X) is achieved in a round-robin fashion using not exists X such that not P(x). For example, The sum of all loan amounts for each branch must be less than the sum of all account balances at the branch. SQL> create assertion sum_constraint check (not exists (select * from branch. where loan.branch_name = branch. branch_name) >= (select sum(amount) from account where loan.branch_name = branch.branch_name)); 3 TRIGGERS © A trigger is a statement asa side effect of a mo ‘executed automatically by the system jon to the database. 1, we must ‘under which the trigger is to be executed. © To design a trigger mechani (@ Specify the cor (b) Specify the actions to be taken when the trigger executed. Database Management Systems 15 ‘© Suppose that instead of allowing negative account balances, the bank deals with overdrafts by (@) Setting the account balance to zero. ‘Creating the loan in the amount of the overdraft. iving this loan a loan number identical to the account number of the overdrawn amount. © Thecor for executing the trigger is an update to the account relation that results in a negative balance value. Example ‘SQL> create trigger overdratt-trigger after update on account referencing new row as nr for each row when nrow.balance <0 begin insert into borrower (select customer_name, account_no from depositor where nrow.account_no = depositor.account, insert into loan values (nrowaccount_no,nrow.branch_name,arow balance); ‘update account set balance = 0 from account where ‘account.account_no = nrow:dccount_no end; - ‘Triggering Events and Actions in SQL. © Triggering event can be insert, delete or update, * Triggers on update can be restricted to specific attributes. Example: create trigger overdraft-trigger after update of balance on account. ‘+ Values of attributes before and after an a update can be referenced. (@) Referencing old row as deletes and updates. (b) Referencing new row as inserts and updates, * Triggers can be activated before an event, which can serve as extra constraints,“erep Sunsrx0 JO vorworpipou jou Inq “Bp Mau Jo UoROsU! sMmolfE suoREZHONINE esU]—*Z “ziep 30 wonwoytpour jou nq “Bupeas sMoype cuonezyoNINe Peay. “T ‘asequiep ox Jo sized wo wopeztioyne Jo smog NOLWZISOHINY $1 ‘sonoeid Sumy yar “s supne wed *y “ajo ‘eBejoqes ‘spuomssed uoqors woyy uonsaIOg —« soy warn “5 “sayeaig Ayynoas ywoisKyd 12a19p 01 sanbiuypay aremyog —& “a1 eatjod ‘SyD0] Aq Ayund9s [RoIskyd . 's9o1A@p Jo sted {Gano uaamiag syped ssaooe ajdinyy “o¥e sossng yonp ‘SysIp paroutur-areaproy uorwondoy 8 10s ‘eBeurep orsdyd ‘Suyddospsanta oquonsay9 aaiseaur WM WONT S9[qUD JOUIULLD pure yoH}aU Jo UONDMIO¥g —« ‘2y9 aBewep jeoisKyd ‘oumses “yarn Woy S¥SIp Jo uoHosIo1g —« “a}o omy zamod ‘spooyy wioyy wousdinbo jo uonsaioig. aay qwoisiqa ‘F AydesBodkig (paseq promssed) joooioid uoneoynuspy = « asm suusyueyo2H ‘safessaw Jo uoneayspour 10 youn Woy parooroud aq isnui SxUYT —« SOUS poisnn yi soreoyunumuoD HOM auNsUO ISN aS YSU feaay yomyon_“f th SWAISKS iuowaTEURYY a50qa00 “suonanuisur auryseu paBajiaysd jo asn zadosduut woxy wonaraig = ‘Ayorpne ,rasnuadns, jo asn sadoudut woyj wopoarorg = poorord s50008 faRa|-oyty® piqeaur woy uomoaorg = # Bupesedg, “ep pasinbor 01 Ajuo ssa00e siasn ayzoads mo][e 01 wis!uoypoW LoHEtIOMNE pure woHeONUAKNY PAs] masks asequied, Ayanoas Jo sjaaa] yuaxoyyyp aij 248 24 -ewp Ay{pout 10 years 01 sidwone snojorfew wow wonsaaxd esti] -siasn payuesun Au ysuree wep atp pronayes 01 yenuasso st 11 asne09q SWC U! \deouoo TuEHOduH st wep Jo KNIIg ALrunoas: “saseo Aureus ty s1988gn Jo puarsu pasn '9q uvo sanyftowy uonensdesug sox ue Buu 1eyp J28u0] Japisu0d uv ‘s10831n Jo suey Yons Jo yBuDq aMp st dK mss asoquiep Y* 2qp yo 19s reIp qwowareIs arepdny tine paysarap 101s 198840 w 2 fapArasuy ayp Jo aunyER axp sesne: 700 1eaU8 UpIAY UMM 2 Plnoys s1283 sua asn 0} Jou ‘swou Jo soquinu 2818], Buyjeap ways yus}oyyo axoul aq UE aropdn ey stuowiowes TOs tht ‘sou paoaye Sum uo9 $9]961 ‘resoduay 0} s9yar 01 [ge Mau BufoUazayad 20 2jqe PIO Suyouasaja1 95 # ‘wor yea 40} Jo pEaqsUl WOUIOIES YORI IOFIS. # ‘uomnonsutes) & £q paroayge savas [78 405 POH! {aq uso uonoe affiuys v Mor paraayge yea Joy uoHe azesods v BuRNONXe Jo PEaIsu] sne@fuy parywam Tapnoeg pur Suva3 4 Integrity and Security Tpdate authorization: allows modification, but not deletion of data Deletion authorization: allows deletion of data. rms of authorization to modify the database schema iew Example Index authorization: allows creation and deletion of indices ‘Suppose a bank clerk needs to kt but is not authorized to see speci Deny direct access to the loan relation, but grant access to the view cu: which consists of only the names of customers and the branches at wi have a loan. ‘The cust_loan view is defined in SQL as follows: SQL>create view cus{_loan as select branch_name, customer_name from borrower, Joan where borrowerloan_number = loan.loan_number; ‘The clerk is authorized to see the result ofthe query. SQL>select * from cust loan; ‘When the query processor translates the result into a query on the actual in the database, we obtain a query on borrower and loan. ‘Authorization must be checked on the clerk's query before query processing begins. Database Management Systems 19 ‘Authorization on Views * Creation of view does not require resources authorization since no real relation being created. > ‘* The creator of a view gets only those privileges that provide no additional authorization beyond that he already had, ‘+ Example, ifereator of view cust-loan had only read authorization on borrower and loan, he gets only read authorization on cust-loan, 7.7 GRANTING OF PRIVILEGES ‘The passage of authorization from one user to another may be represented by an authorized graph, * The nodes of this graph are the users. * “The roots of the graph is the database ‘+ Consider graph for update authorizat © Anedge U; U, indicates that user U; has granted update authorization on loan wy, us S u DBA < u— us uw Fig. 7.1 Authorized Graph Authorization Grant Graph © Requirement: All edges in an authorization graph must be part of some path originating with the DBA. * IEDBA revokes grant from U,: (2) Grant must be revoked from U,, sit (b) Grant must be revoked from from DBA through U, U, no longer haé authorization, 1, since U, has another authorization path“Yoon asuqewep ofButsw 0} paddeui ag feu de we jo siosn-pua fy ‘siansos uonvondde Woy 11 ay Hd [Te ayoaas 0} [e 99 Kew 4 “Pozinbararv sayoau Suypeoseo 3 stay pueuswioo ayOnad atp OINSOs IAL ‘pusses Er) Zp ‘Ip woul young uo ways ayous1 wouf < upuaate 10 auous NO]04 > wo < 184) 28% > ayou24<108. “uonezizoqne axons 0: pasn st yuowete%s oonal OH, TOS W uopEZHoMny SupyoAsy ‘4aBupW 0} wn0200 wo saBapisd jo puns saSounu aos o282<105 ‘9101 49410 01 uono pure ‘siosn oy pauBsse aq ueo Sojoy . — “asm 24 In ‘sajou Woy Payonas 10 0} paruEs8 aq Wed SOBOTEALL Surpuodsaiioo v Suyeao £q aot0 ‘soqto 01 289114 nudo 1wod8 yt I) 01 Younag wo yazas puvs8 < 70S dues 01 '7 swore pus young wo se8oj aud sooos ain 9+ ord sou 2}qeaoy p18 205s} voys wae posn :saBoAtEd Ie -uywop paytseds wasn 0 ssn & sazuowne 6-108 Ut Bsn ££A Ep ‘Tn 01 young wo poajas mes8<10§ onbjas youesg 247 wo uonezyso¥nMe 199)98 ‘7 pure ‘7 “9 sso8n wR opdurE "mats ayy Sussn Aionb oy Auge 94920 “Uo}ej4 0) s8a5B8 peBd SMO 139195 Tos uy seBey Aug (va 9p 29 20) “wart poutoeds omy uo afoqraud am ploy Speome snus aBojitad oo sow oy], 6 stonejas @undyzapun amp uo sa@ajtAud Sue Sunubi3 Xjdut you saop mata e uo aBoqlaRd v BuNULID aBoqtarad ayy siosn pra ype smo} << 2ubU AafA 40 atuDU UO}O|aA > UO < 11) aBaq!AI4d > JUMEB R. ‘the value of B is unknown as is, the comparison A +B evaluates The concept of nulls, at least as 4 logic in which there arc three .) Hence the term three-valued that term is usually understood, inevitably leads us i ‘ruth values [true, false, and unknown}(an) d YO" 4. #01 uayeatnbo oq or pauyap st (4) A sisixa J J uoyssaudxo oy) uo 2}govs aay w se sino A YoIY uy woIssandxo UEJO09 va 1 a (404 () pv sno sofas yay 2pqeutena8uE1 ta (Q) sojdey tpy wonejaresy 2.2 ‘spton s24n0 uy -fonnoedsat “GNY pue yo pateiat se siayruenb asain aUap 94 u : “TIVUOS Pue LSTXA 10} TA¢ Jo suoneordun ayp saprsuod 0} pao ‘op om ‘sajdurexa no jo ~~ aaAVW 'Sour 9ys186q® Sv snynoqeo tp ue Joye! Baqadeaxp as aM zoND19Njaxpaids9q sSmolfog SP 2IQR) YIN YIM “AEAWI St 9u0 yweyodur oqioue ‘1a4aMo4 ‘paaut aM suoteiado ue9[009 Kuo ay 0U 388 LON PUY YO “ANY wun :(9= V) LON : enn: 9>@yoa> Vv yun: 9 <@yod AUVIVS 1L61,) ALVA > Oa * NV (8 NY .touruesBoxg, = €0f ) LON NV Suoyssaudxo Buyo|fo} 241 aH, “NASI pue'p—¢ *¢ — V asaddns ‘ojdurex9 104 ( ANWIVS) anv yng | ’ ny | ; (40d) YNASI-¥O naa]on gan (8I-I-1.61,) ava > aoa) any papa seed 3 (Ca0r) ANA’ SI YO ga | ao gay | anv souueiiod ,=€Of) — AYTHA\ AAO jusqawos Yoo} 0} sey Axon ‘Toramoy oreiado o¢n mouiyy (Yoanoadsau “Ty; Pur “ava "VHD sadQjo AIA TAIN 18419430 AUV'TVS PuE*GOC “AOL somnquire Mp pauunsse ancy a9) Ponyes-aan yeap are ‘3 ew sousy pur ‘2s]aj 30 anu jo peau “anjea yin uovourxun xp 0} SapEn|BA9 NIT S} (00'00005 > AUVIVS Spuesedwios aup jo soya yoryas ut uosyedios sejeos Aur yey) pres Apeaure aAey aA, NV (.81-1-1461,) atv > aoa Ny seunuesdord, = qOr) ACAVW PIA TIAOTIWA s1oye19dg av2joog, ‘uo}ssaudx9 un2|009 Wo: Ajareipawun suonaasqns om) Yoooudde Ip¢ ayp Jo stuuoduion HOVOUddY TAE SHI JO MAINNAAO NV_2'8 Voneuaoyay BUSS wt joads SyINA Hujuwaw-siim Jo roa ayy (BtIMOTIOR e ‘Swiss juawuedeuBMissing Information and r2 (of the same type)to be that rel ofall possible tuples (such that tis a both). Update Operators mnt (again of the same type whose body consists jeate of some tuple of rl or of some tuple of r2 (op smpt to place a tuple in R either INSERT or UPDATE in ich the A postion is UNK is an error INSERT or UPDATE to place a dup jon of “duplicate tuples” here ‘An attempt via eith error, as usual. The dé subsection. Integrity Constraints ‘An integrity consiraint can be regarded, loosely, as a boolean expression that must not evaluate to false. Not, therefore, of a WHERE clause, soit is regarded (speaking some what lowsely in both cases). 8.3 SOME CONSEQUENCES OF THE FOREGOING SCHEME ‘The 3VL approath as describe: consequences, not all ofwhich are imm of those consequences aad discuss t! Expression Transformations First, we observe that several expressions that always evaluate to true in 2VL do ‘not necessarily always evaluate to true in 3VL. Here are some examples, with commentary. Please note that the listis nowhere near exhaustive, for the purposes of an integrity constraint the previous section has a number of logical ‘obvious, In this section we point out some significance, * The comparisonx = x does not necessarily give true, In 2VL, any vale x is always equal to itself. In 3VL, however, x is not equal to itself if it happems to be UNK. ‘The Departments-and-Employees Examp! ‘example in some de in Fig. 8.1 and the expression here are implicit range variables. For the only tuples that exist in the database, this express evaluates to unk AND unk tl the expression Append an additional restriction term a= et Management Systems 87 ‘The boolean expression p OR NOT (p) does not necessarily give rue In2VL, the expression p OR NOT(p), where p in tum is some boolean expression always-evaluates to true. In 3VL, however, if p happens to evaluate to unk, th overall expression evaluates to unk OR NOT (unk), that reduces to unk. not true. properly of 3VL, which suppliers in Bombe and take the union of the two results, we do not necessarily get all suppliers. Instead we need to include “all suppliers who may be in Bombay.” (In other words, ar expression that does always evaluate to true in 3VL-that is, the 3VL analog of the 2VL expression p OR NOT(p}—is p OR NOT(p) OR MAYBE (p),) The expression r JOIN r does not necessarily given r strate as follows: If we execute the query et al ” followed by the query “Get all suppliers not in Bombay, In2VL, forming the join ofa relation r + ((e, join is idempotent}. In 3VL, will notjointo itself, because j testing, not “duplicate-style” eq INTERSECT is no longer a special case of JOIN. This fact is a consequence of the fact that (agai equality resting, i In order to illustrate the problem of incorrect transformations, we discuss a specific i Consider the simple departments-and-employees database shown DEPARTMENT. DEPARTMENT_NUMBER = EMPLOYEE.DEPARTMENT_NUMBER AND EMPLOYEE . DEPARTMENT_NUMBER = DEPARTMENT. NUMBER (‘DI’) (which might he part of a query, for example); DEPARTMENT and EMPLOYEE is, tounk. However, a “good” optimizer of the form a=b AND b=, will therefore infer that a=, and hence will iginal expression DEPARTMENT.DEPARTMENT_NUMBER = EMPLOYEE , DEPARTMENT. NUMBER ‘AND EMPLOYEE.DEFARTMENT. NUMBER = DEPARTMENT. NUMBER ( AND DEPARTMENT. DEPARTMENT. NUMBER = DEPARTMENT. NUMBER.J9MBo Yowa,Jo sowordnp axe Zi paw» a=ne ‘juayearnbo aze syuautayeis Suotty op yyy esto og safuoy ou mou sym ‘taADNOY ‘ION ‘poBuoyoUN sordde lwonoofoxd Jo uonuyap yeurdt40 amp ,‘sajdny ‘yeo4[dnp,, Jo wonruyap papuarxe sien TIA "waouy Uou St X Jo anfea 2x7 TeYp st INA St X,, JwOWO}EY yp ‘seu0o Aq “yun 9g 0} umouy 5 uoulayes oun ‘sy “yum 10 ‘a5 * ‘adki Jo ajgeuea v st x asoddng SEN Uy Jo onyea amp samp = Suues otf aazy Aatp jt ATUO pur 31 Jato Yous, ‘ou 2051 3da008 Mp JauR2yN eBpaf 0} NOK oF 31 axBOF OM 1 2 Stang gum wa1apEp ¥ dope 0} ajqissod sy aoe 19H “SUONIPUCD [eAdtnal Jo worENTeAd ouput Busy Aaytenbo werp teap Jo jaa] zamo UNION ST IND St [Baowtas ayeo1/dnp 103 [Bupsay Aypenba) sis: (1 <2 A) aSAVW) ATIVYOL ssmorloy " se Keme poureydxo st hofpenuoo jwaredds ayy ‘uorreuruye ajdma ayeonjdnp Jo sosodind ayn: (1<@:0) ATIVYOd 4a} s2t0 Y>RD Jo >», 24 0} Padapisuod [INS axe axou ato o} Tenba Jou are foxy asiay (1Aa ‘Sy¥NA om] Zou st uoNsonb ‘SIU 01 Jomsue aqp “ppod 01 Supiosoy on: (OW wNOT SD ASSIA 31°S1} Jo uoAe rou “Bury yo oyway{dnp v aq 29A9u UO NIA te suIEHUOD eM : < 2idma & yey opnjoucs 0} poaioy ua am 1aa9 10u “BunypAue 0} yonba you s} (u30s an (E IS [NOT] NULL red type (and here it makes no lar or nonsealar). To fix our ideas, let pe. Then there is clearly al V having at least one component ‘ull even ifall of its components h the standard is unclear on sue) that its components are considered to be null as well. Thus, if V is not m least one n the comparison V = V gives null, and yet xpression V IS NUI we can at least say V=V)ISNOTTRI 1 or it has a null compos Ifthe “constructs” a row of degree one, then SQL treats as if it actually denoted the value contained in th Towas such; otherwise, it treats it as denot ; then the expression r IS NOT NULL the first is equivalent to cl IS NOT ond is eq tase Tables OR c2 IS NOT NULL. Another consequence is that some nonnull components, then r itself is apparently neither null nor nonnull Columns in base tl associated default value, and that default . ponent then sels apparently neither nal eno alue is frequently defir to be null. Furthermore, columns in © Tests for true, false, unknown: Ifp is a boolean expression in parentheses (actually ase tables always per constraint—usually just NOT the parentheses are sometimes unnecessery, but they are never wrong), then the TULL—for the colum: them, following are also boolean expressions; are (o be true to our own principles, we pIS [NOT] TRUE ther e PIS [NOT] FALSE PIS [NOT] UNKNOWN,“enn suman ‘astauaryo (©) 24H fe 49 Jp yey suman 39 WOW! JONLLSH S1y97 ‘a yroq axe ry pu r7 (q) 10 ann 24 3p Yons 24 psn Py pu y 710 stuauodusos qu SyBry pue yr] 24 uonsant 04 OUP 127] TaUO Ye Jo “ome suony S2jun uays ana sumnas sauunawios andINA “x0jaH0 351 puv ‘ann saai8 zx 1x uostavduioa ay) 18) 04 sureiu0a Increased. Parallelism: Queries on 7 ‘may be processed by several nodes. ‘in parallel,f TERS fy tomeyas jo aydim yoeg Ma Za st 4, uongjad & ‘uoneuawie, ns Jo oquinu 8 ow pouonned ‘Mone iuoMBesy TeyWOZLION (4) “aus 24 §89008 0} a|qe are ays ou0 8 srasn ‘ways paingu ee 1 "suoseay Aiyinoas 40 ules soBeyaeapy IMarayp ve poroys 2q Kew swuouiBey soy “erep axp Jo woHe09} Fonqndoakojdua +z [roisdyd ax sou 0} pasmnbax you az sias() :Xouauedsuwsy, woryes07y (0) = ojur awa s06ojduso -pooeid wo0q oney svordax azoqy's 10 “suonpjas out poruawsseyy aq Aeus uonjau stu ‘suoswas Kowatsd 104 ‘Parmoriday uoaq envy si99%go erep reyNA MOLT O1 partnbat you are srasn ot, “anbmun jjeo180] se yoo(qo ewep yoes MaIA Stos/) :Aouauedsuvay woHEoydoy (4) Pomouiexy uo9q seq NON aX © Moy wou oF posinbas jou oe sios7 :Aousuedsueay, uoReUoMeLY (8) neu isap ‘ouren ‘pr-vaXojdusa ‘2asoydw peo 103 sasoys WIN JUL 90Ko}dwio uorBfaxw YL aseqeIOP AyKEATUN € JOptsUO) adurex: “aa Da Dd ed cor ‘of yeanyeu axp Surye) Uds . Zyumo00" (qunos9e) ,opisttH, = PU YOURS . jyuo09e ssw MAIEBTP omy om axon Uatp ‘aBpUALDY pu apIsITE-soyouG om KUO soy Masks BuO 1931 Yotg rman ed & 0 BuxBu0}9q stuno9de Jo sajdny Jo SSIsUOD KOT Jo Yea ‘SIUAUIBR WARBEP [ea49s ow poprarp sf uoneja Kmoooe xp ‘uoREMaUUEY eIUOZLOY Huts) onenwameyy eara, gy sadi of, WORE tous urEWuoD syuoweAy 9594, ‘powuauuiieg st 1, uONe|a J vopmuewBe1j Beg ZZ “yonuo Kowaxinouos Jo Kirxajdusoa pasearouy dau Sere}409 AOU. SSMU 19 ane 4 woRIE|92 8 JO jepdn wo peoysoao pasvasouy soBeyuwapesy 88TH Ap ES J16q] JeATag/juaK| pure SasequIEG poInqiasiC| 0) VOIRpOAY ¥ (coueyeq ‘ours young ‘ou 900) = wunosoy : “uoneyar yunosae ayy z9prsuoy Rurfyo axp jo uononnsuosar moje oF uoEULX ajdaex . tr" 2a La siuouseu Jo Joqumu B oR pop “amy Up lena IYL “stuowoy jp Jo uorun axp Bupqes Aq 4 worear xp ronuysuoDay aM, OMonu ‘eow2Ko poseasour st Suse 01 paweBedoxd aq snu atepdn cu worweindtuoa smoottuta pe yup amsua yas woke ou "a quautBe ionnsuoo ot pasn st iq avworpaxd ayy, *4, uonwyar Teqo[@ 241 uo roajas v sv pauyap aq uvo juoUBeYy feIUOZLOY v ‘jeLauas Up ‘Popsou J ‘poronnsuoday 29 {weo UOHE|as feurBtI0 amp Hp os ‘siuoBey oup Jo at0 sea} eo} BuOyOq ISTE 56 SuaIsks WOURaeUEbe ablew continue operation. Disadvantages Software development cost: The implementation of distributed system is, ‘additional computation re of overtead, that does no 9.3 HETEROGENEOUS DISTRIBUTED DATABASES * (b) Organiaatic from a variety of pre-existing Many database applications require d: i tion of hardware and software databases located in heterogeneous col platforms. Each local database management system may use a different data model. Some may employ relational model, whereas others may employ network or hierarchical mode! Transation commit protocols may be incompatible. Concumency control may be based on different techniques locking time stamping ete. “Manipalation of information located in a heterogeneous distributed database requiresan additional software layer on top of existing database system. This satware layer is called multidatabase system. Fall integration of heterogeneous systems into a homogeneous database is oftendifficult or impossible. - The investment in application programs based on cexisting database systems may be huge, and the cost of conversion may be prohiiive. ‘may notbe po to corpuration or organizations. Database Management Systems 97 93.1 Query Processing Query processing in heterogeneous database can be complicated. Some of th issues are 1. Given aquery on a global schema, the query may have to be into queries: local schemas at each of the sites where the query has to be executed query results have to be transiated back into the global schema. ified by writing wrappers for each data source, whic provide a view of the local data in the global schema, 2. Some data sources may provide on} they may support selections, but : For instane yeed to be accessed to answer a give ‘may have to be processed to remo ince the query executi query plans at differe ization, and just use heuristi heterogeneous data source: very facilit tor systems do not bot ‘The term virtual database is used to refer to multidatabase/mediator systems since they provide the appearance of a single database with a global schema, although dat exist on multiple sites in local schemas, 94. CLIENT - SERVER SYSTEMS AAs personal computers became faster, more powerful, and cheaper, there was cl system architecture. Personal computers supplanter Functionality provided by database systems can be broadly divided into two part: ~the front end and the back end. The back end manages access structures, query evaluctior202 Avy *p 9007 Aeyy] [9] “ojdurexo yy uorejar Iuptowie 205 soyovoudde snouwa oxp uejdg gsuombyuauiBey eep am yA °9 007 EW ® L007 .AON ‘S007 «AON [8] ‘sosnqelep pamngunsip wo sou payreap B aL °S oz dey] ‘asuqeiep pemnqinstp Jo uBtsop amp urE|dxg “p 07 AON £007 AON] “wreidxg csaseqeiep panqunsip Sunuawo}dun [rym sonsst o1seq yp are UN“ OC ey] If] swouuostauo pomquasip» sad omy z ‘aqdurexa aq a> Lava 2 &eWi] worsks pomqunsip eyo aunioonyare axpayeorput or YoIoys WU VARIG “OL “Soseqeiep pangunsip jo saBeweapesip arp uonuayy “6 “wonsesuen joqoy8 pue uonoesuesy [e907] auyod “g IZ AON) Murex2 qu suonmuawiey snows ay aig gsuonEINaUiBeN eep ame THA, *L Avy) ‘woneoridas wep Jo soSueapesip pue soBeiueape op aIag “9 Z 40] sosequiup pamnqunsip 'NOr2J91 (HUW saseqerep snosuaBors19}] puv snousBowoH] ayenuaIayiq ZAeW] -osequrep painqunsip jo woudojonap ain 10 suosvas ayy 3n0 11] 24H 8 Fo0z 22a] 1 key) ¢Sseaewp parnquisip sup uy uonejanw azors oy payovordde ow) amp are TEAL. casequiep painquisip Jo soBeuwape atp axe ey AON] "oisis aseqerep ajqeteae Sterosouoo au0 Au uoydxa KyauE, * V- Lava SNOILSAND M3IAaY ee ‘swaishg quouaTumyy aroquied, ‘sfe9 aimpaooud ajowras jenpyarput 91p JO sioayjo 947 opun wea JoAJ2s amp ‘sHoqe wonsestEN 2ip J ‘smi ‘pua saAzOs axp Te UONOBStEN, [Buts © ut pasojous are 1 ‘zaumuresBoud amp ob YP AUTOS 0} 90u}IO1UL 2SeqereP 219 0} JUaYO v Sw SIoe oY UT ‘aAIas uONPoI|dde oy}, saKIas UoHEodde Uw OF Se Yoru 1asaorq qaq BSI pud yuody atp axoy ‘(] JardeyD) {| amy UI Jo4|IeO Aes am {gorges ‘aumsonrgore san-sanyy e dope siosn jo soquinu 9810] yas [eap eyp SwOISAS, is spuo ywoay apraoid fou itt 29AI08- WHATIO ax as ‘ypaya Ul ToAI2s pus-yovq v Woy BiEp ss90o8 0} Af ‘safeyoed stsyeuv-jeonstreis pue sjaayspeauds se yon: eo eoeHamu Saar 10 Ha podojaaap azom ‘¢ ydoy5 ut Aanouotounf pus 209 pu pus nosy £6 Rg out peg 210s (av+ 19) a0np.H0 aomjiont | | oomponuy pu wo suuoy || sosn-70s "sun rep has warsds sasaas - mato » fo aunonus jousuay 26 By l youu 1 r won] [Ton Idde ue ydnosyp 40 “JOS ySnomp mn B8sn TOS se ons 5001 Jo sis}SU0 ouauinauoa ‘uonezrumdo pure ‘Waits asugeimp v Jo puo quo ayy, Clanooar pun “onuod fouaunauoo ‘uonezumdo pUE SStegreq sugar pu sseqsINC pomnguic| ay voRDnpanN v6BBBBBBB Unir-3 Dataase Design Cuaprer - 10 Reationat Database DEsiGn Introduction Functional Dependencies Functional Dependency Theory. Normalization Multivalued Dependencies and Fourth Normal Form Join Dependencies and Fifth Normal Form Domain key Normal Form (OKNF)“aa|goxd Aoaopunpor ¥ 0} spea] sry. ‘sou 7 pawuosaxdar ame syreiop wMoHUHOg YouRNg 12}] ose. eT worpey_ ost sie satey 002. er a mou | oumueo; | saoysno | sasse (runow ‘ou-uvoy ‘surou-s2woysns ‘srasso ‘Gyo younsg ‘sumou-younsg) muayas-SuipusT BUDS uonEpaL oxp Z9pIsIO “squpensuoo Ayps8oquy aseqeyep Jo UoHBjorA 104 sayepdn Jo BuPyoayo orp wee “pauosoudas ame somqinye Suoure sdrysuonoyar yerp amsiig “Tyep juepunpas Prony . a “uonBULosT uTET20 ywasaadas oy AUpIqeut (q) ‘su|goid uonepdn ‘uonajap ‘uoniasut 0 speay ye - uoreuo;Ut jo uontodoy (8) 1 peoy Aru usp peg y udysoq esequieg jeuonepay ut syey-d UWoHE[a1 Jo Wonsa[[o9 ,poo8,, w puy am yyy sazInbax UBisep osequiep jeuoHEIoAy NOILONGOMLNI 1 NOISA ISVAVLV TVNOLLV TA 0 YaLdVHO102 Relational Database Design Redundancy Data for branch name, branch_city, assets are repeated for each loan that a branch makes: (@) wastes space (b) complicates updating, introducing inconsistency of assets value. Null Values (@ Cannot store information about a branch if no loan exist. (®) Can use null values, but they are difficult to handle, Decomposition ing schema into branch_city, assets) 1_No, branch_name, amount) Branch-schema (branch _ Loan-sehema (customer_nam All attributes of original schema (R) must appear in decomposition (Ry, 29) R=R OR Lossless join decomposition. Alll possible relations r on schema R. ©) Dd me Goal: To decide whether a particular relation R isin ‘good’ form, decompose it into a set of relations (R}, R2, » Ry) such that cach relation ‘the decomposition should be lossless decomposition, 10.2 FUNCTIONAL DEPENDENCIES _Ttrequires thatthe value fora certain set of attributes determines uniquely the value for another set of attributes. In given relation R, X and Y are attributes, Attributes Y is functionally dependent on attribute X if each value of X determines exactly one value of ¥, which is represented as XoY ive, “X determines Y” or “Y is functionally dependent on X” X-— Y does not imply Y > X 103 the value of an attribute “Marks” is known then the ‘Grade” is determined since Marks + Grade. (@) Full functional dependency Ina relation R, X and Y are attributes. X functionally determines Y. Subset of X nally determine Y, [Sais } <>) [oor pt In the above example marks is fully functionally dependent on student no and ‘urse_no together and not on subset of {student_no, course no} This means marks cannot be determined either by student_no or course_no alone. It ‘an be determined only using student_no and course_no together. Hence marks is fully functionally dependent on {student_no, course_no). {@) Partial Dependencies Attribute Ys partially dependent on the attibute X only ifitis dependent on a subset uctor_name are partially dependent of composite 10} because course_no alone defines course name, (©) Transitive Dependencies X, Yand Z are 3 attributes in the relation R. x 3 Y Y>uz xX 32 >Peplone oq pinoys ajqe ur sutunjoo jo sanju Jo uonnedas ayy « “Paprone ag pinoys sojge ur sutanqoo a1gaymaN,« “au Jo ad apts soy ep azors pmnoys a1quy yoo « ‘voneja4 B Jo 2oueistY jpe 49 paysnes st auDu-rowoysn> ou-uwoy ‘Gens soungo Aq Sous vusayos-ueo} Jo ours ag}00dsv‘oyduroxa 104 “s2oueysut [#89 [1@ uo ploy tou soop Kouopuadp fo Aouapuadap jouonouny v Aysies ysnuu euTatOs Worx v Jo 9 ‘suomejas [eB9] Jo 428 94 Uo st as 10 es om ‘soyouopuodap jouonuny Jo 4 195 8 sopun 88) st woe “sefouapuedap etonouny 0195 ton wpm feo ame Koy s1296 0, SuMNEFANISAL, gp © sejouapusdop feuonouny asn 994 soppuapuodag jemopoung jo 267) ¥ © vy > vous0j puwey — y J So paw jy 303 £¥ arepmpuws e sty — & YH Suo puw 31 y wurayps uowsjas 10g Kowadns wsty ge Plot S90P V €13nq Plou 0H seop gy ‘souRIsHL StI UE ‘waypuopl us axvy pmnoys 2qqe1 Yow 9m Btsap asoqeiep poo8 w anaryoe 01 among 29 plnons iin sans wog “eseqeiep jeumndo we 30 St -torsts amp yo aouo}iod ax ss uuonworidde tm Jo soueuns0js0d PeUWON “aqRIOHaeP Ae asoqEp 9f Jo zaquume atp 3 “toy uioo pue zaquinu ax 19A9MoF] TIAN 95 “2iger Jad soxopur may 4 1youiaq snosount Set] uoezKyeKHZON ‘onepunoy ay sky asequyep 2tp o ut {isop sood ¥ -uB\sap aseq eiep atp 1 port HL ‘So7ge Jo omionsys ayy Ayr 24N Jo coueunojiad ax ‘ase WF oswaxout oun ya aseatoUT 5 2% Jo ssomseduioa ox ur we p X9pu1 paw Sunzos s2}sej apnjour asay, Lye “Panttno am suonejar saxpo Ut 23 SIH “Pat2qep 40 poypdn Buroq ayqen ato Ut EIEp Jo pooysa: yo Buypuodsou09 vit +4 Jo soueysut Burmoyjoy 24p (2'V)4 9p Isu0y sydwexg = (6) [oy = fot sr req, i ‘d sonquie oy uo aorBe Jee ou Wo oaxfe 4 Jo % pu 'y sojdny 7 Atte sonauayys *(y). 10431 uo puv jt wo spjog go Aouapuadap euonoury a4] gp US dpmwy so oq, ‘saimqune Jo “Ag % stonay yaa ayy ‘waray>s uonejaxe9q YIP] Ssneo eur top Sanojap puw Buréyspow Bupiasuy - s ‘Pareadax oz stony vaya ano90 o} fx] azoUs are S101 - a TP Sony 83 wonefax pe £01 be9] uso Kouepunpayy Couepunpas Dept. S value of DeptHead. Hence Dept: Hence all the attributes are function table isin 2NF. th is not a primary key. Guidelines for Converting a Table to 3NF Find and remove non-key attributes that are funct attributes that are not the primary key. Place them i dependent on the ifferent table. Group the remaining attributes. 7. convert the table employee into 3NF, you must remove the column Depttlead since itis not functionally ‘dependent on only: the primary key ECode and place it in another able called Department along with the attribute dependent on. Umployee Feats [oan Eto [Sens |E305 [Finance e102 [sae e508 [aan E —Fanaee E08 [Fame | will have to be made in every row of the table, otherwi 14.4 Boyce-Codd Normal Form The original det isfactory for the tables: + That had me ition of 3NF was inadequate in some situation. It was not Hence, a new normal form - the Boyce-Codd normal form was introduced, You Understand that in table were the above three condition do not apply, you can stop at the third normal form. In such cases, the third NF is the same as the Boyce-Codd non eal ‘A relation isin the Boyce-Codd normal form (BCNF) if and only ifevery determinant isa candidate key. Consider the table Project given below. Project fECode | Name [ ProjCode] Hows Et Veronica | P2 8 2 ‘Anthony | PS 100 es Mac | PS 5 lea Susan | PB 350 lea ‘Suan | P5 5 a Veronica | PS «@ This table has redundancies. Ifthe name of an inged, the change inconsistencies.Ur Soinqunse amp yeep Yons ,. UL A X Soxouepuadop Jo 198 ou StF JO 1989 (Dre porous ‘ra uo 4 Jo wonsefoud amp “yuo J sexoUapuiadop Jo 1s © UOAIE) ‘uontsodurooap jo uopenleseid Kouepuedeg Ipod wonwaxosaid anqunye payed. ua aD EWLIOY 1S0] are soInquME ou yw os a ona “WoRptiog uoeaLosaud L LK} Jou ao Rou BooL null 12 KL Ki KI K2 Consider the mlation described in table Relational Database Design Hence the praction of F on each relation schema R, in the decomposition D is the set of FDs in F*, suchthat all their left and RHS attributes are in Ri. A decomposition D={R,, R,...R.} of Ris dependency preserving with respect tk if the union of the praperties of F on each R, on D is equivalent to F. base Management Systems ‘A schema that is in 3NF but not in BCNF has the problems of * repetition of information (¢g., the relationship values (¢.., to respect the relati where there is no corresponding value for J). 0.5 MULTIVALUED DEPENDENCIES AND FOURTH NORMAL FORM 0.5.1 Formal Definition of Multivalued Dependency and y are both subsets of R, specifies the following constraint on any re and t, exist in r such that t, following properties. Here, X > Y denotes that x multi-determines Y. An MVD X —> Y in ris called a trivial MVD if - 8) Yisa subset of X or b) XUY=R Eg. Fig. 10.1 (b) ‘An MVD that satisfies neither (a) nor (b) is called a nontrivial MVD. 10.5.2 Inference Rules for Functional and Multivalued Dependencies IRI (reflexive rule for FDs) 1X 2Y, thenx sy IR2 (augmentation rule for FDs) {xs y}Fxz—s yz IRS (transitive rule for FDs) f {kK >yy— sz} Ex sz IR4 (complementation rule for MVDs) (> y}F & > R-XUY)} IRS (augmentation rule for MVDs) IEX-SY, and w > z then wx > yz IR6 (transitive rule for MVDs) If (XK DY, Y-9Z} EX (Z-¥) IR7 (replication rule for Fd to MVD) (XY. Expy) A multivalued dependency (MVD) x y specified on relation schema R, whe state rof then two tuples ¢, and0 SLO - aI suortmpa4 a1Np 048 1 dN Russoduooacy (9 AINYN & INENG Pur IPN © AVN SCAN 0W1 116 uonmas GAT AY (0 wisof pouisou yranogy [9] 2130 somann sos m9 Jo stasqnsauw 'y nay “y “Py jo wot dap oF enbo ta tuo pu 3 Cy gy) Souapuadap utofsousies x won y Joes yesaags tous pojonnsto9 2q ou AN Ut wonofas y soyRy 29 JomuR® Wp SUORE|eX aNeY OS uuOS fULOU IAF JO UM SUL “CLAW at Jo uowesteioue8 © S| Yon soruapuadop-wot us sfeop WOf[BuNION oe ANS omp Bus suoneyax eam a ae acant tt Woneax ax asodumosap 0} ayqissod aq deus 4 S950 yans Uf “stzofgand Cet Ga ‘ corel Pee nny gtOHeI Om UF pasodonep aq jute ing to asmes0q soqeuoue agepdn wefed | wera x — een QgUE Wepanpar ayF swoqgoud axey weD toRAX v‘SosE0 ame aUHOS Ul WHO} {rinou ath Josiuauayyabes 21200} suonas om pasodttodop2q to eUON won HONG] BANG awed | nw | SH) WHtou3E Y wonax uoatd ayy sem pasinbax sy os passnosip suzy feusOU ot SLNAGNadaa-aWa sLoaroudana WYOs TWINUON Hild CNY S3IONZQN3daq NiOr 9'0 5 wher] A woysseycy 4 doapnig | x uysseyc, 44 0X) pa (4b) seus cons ona 6g qu D sone fered] ok] mem AND Serepora we. UE A << x GAA [eatuON v pur SNP UF10U st emp C| UH D wUIOYps HonEIas ¥ asooys AN YN © } aa OP AN Wr 10U st you C[UL Q BUTaYDS uorREpaNw st ouOKA aTYAN “Z “SINAONaAAM dA 40 SL9ALON esneooq “ANY Ul a0 SININAATC - dV PUL SDALOUd - AN HE MOUs ‘SLNACINAGEA - dN PUP SLIILOUd ~ AWA OW! ATA asoduro09p Jo Aaxiadns w10u St AWWNGl‘SINWNG © aWYNGI Pa AWVNd < aWVNA uot ox tH asnED9q ANP UL YOU St (8) LOL "BEA Jo UMEjAL aH “30 Sawzadns & St x ‘ag & € x fouapuodop LA]uOU 1949 305‘ (Sorouopuadap panfeann puesDroapuadap euoNoUTy Sopnyour yup) q Salouapuadap Jo 2s ¥ or oadsau ypNs JN UES! BwYs UOHRE|an y uulod RULON YNOY ¢°5°0L uuuuobry CWE AU wE [Wo Pue311J0 worssodwooap wot ss2jssoyv wos °y pur 'y souoyos Uone|aL aH 17 fusdos, 1 Auadoud amp q uaa se ‘Spiadosd wo! ss0j soy axp sey 304) ps v Buysoduooap 40} uowIpuoD uororgs puw AxessadaU v st vend wots] a sey wnisoduooep tna tpintimh 4 nveeo poe = w= "a pur (AX) = ‘uroqas onjar # asodttiogap amt 29,3494 A, 2 xuop ZEK (0 puez<—m (@ Ada st £U ay (@ wemp sapiadoud SA spas m sistxo atoup puw & x jT_+ (SCAB pu sci. ams 2ou9897800) ST tlsog asequeg TeuOHE|Ry sro SUOBEIOY INY OU] UOR!sodu1090q WIOP Sse} $807 p'5'0b ror ‘SWSISAS HSUIGBUEYY os0qEtal Database De -INE) if for al 3¢ Management Systems 0.19 A relation is said to be in DKNF if all constraints and dependencies that should on the relation can be enforced simply by enforcing the domain constraint and key straint on the relation, one of R, is R) (b) Every R, isa candi Fora relation in DKNF, it becomes very straight forward to enforce all data base ints by simply checking that each attribute value in a tuple is of the appropriate ain and that every key constraint is enforced. ‘An example’ of SNF can be provided by the example below that d Jepartments, subjects and students ‘The above relations: which are taken by a as all students enrolled formation, ‘The above relation does not show MVDs since th ute not independent; they are related to each other (dept, subject), and (dept, student) without loosing some important information. The relation can however be ecomposed in the following three relations (subject, student) and now it can be shown that this decomp 10.7 Domain key normal form (DKNF) The idea behind DKNF is to speci imate normal form” that takes into account all possible types of dependencies and constraints.“ojdurexo quia Woy jeunou yyy, Pue Aouapuadap wor ayy ureldxg “9] “etdurexs yyis uo} jouuiou ypmoy pue Kouspuadap ponye (e002 » 9902 eva] [g]) ‘oidwexo qm uonezifeuON 20 Pes maagip amp trerdxq wonezreuo¥yuonssoduiosop Jo jeo8 ayp 2Ie§.“L [so0z.sond [8] ‘ajdurexo SP it INO SUHF9C| poo paroprsuod ApeouaH ANE ULSI eM UoReTaLESt kay. 9 [e00z.4onI [8] RIPE SSPQHIEP [BON 105 UL} INE J9K0 INE Jo SoBeIueApe om aBe FEY “¢ ANE snsi9A YN aredw03 “OT 4 104 suuzoy youLIOU snoLeA amp eo susinks wo usdoyy qed {¢007.40NI [8] uofje Aressooou urejdx {sias omgume jo amsopsaptsoud oy 0H] ped snauea amp sensi] Suysn uBtsap aseqerep ouonjare [2002.40N] [p] ‘aIdurexe qm sorouepuadap jeuonoury Jo sidzouos msoq ayy ured“ a- Lava loroz Sew] Quoneztpeunou 4 yea st yey A, {Aouspusdap-uios hq yrwauu St yey AA * “INS euyeq “AND OUNjOC uonejar 8 asodwooap 01 Kaessavou st 1E AYA (eooz ew] quoneZHfeULIOU 10} POOH OMp SE IEYLAL [Looz Xe 7 9002.40N ‘9007 amp wos Simms ‘Boy amp yBNowA preawioy Yores “¢ "SH OGETA om pu Is] CGN ayp ‘suonoesUEN Jo sist] om EM AIG. “| {SL-CL 89d jo suonoesuen [ye