Dünaamilist andmebaasiskeemi nõudvate rakenduste andmemudelite esitamise võimalused ja iseärasused
Failid
Kuupäev
2013
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Tartu Ülikool
Abstrakt
Kokkuvõte
Antud bakalaureusetöö eesmärk oli uurida dünaamilisi andmemudeleid, nende esitamise viise ja jõudlust erinevate operatsioonide korral. Enamik tänapäeva rakendusi kasutavad andmebaase ning vahest ei ole võimalik teabe struktuuri defineerida, siis tuleb kasutada mõnda dünaamilist mudelit.
Võrdlusse võeti seminormaliseeritud ja olem-tunnus-väärtus andmemudel. Arutati nende olemuse üle, seletati lahti info paigutus nendes ning testiti jõudlust erineva suurusega kogude ja koormuste puhul. Töö viimases osas toodi võrdlusesse andmemudelite operatsioonide ajad ning tuvastati andmesisestuse ja -tagastuse päringute jaoks parimad mudelid.
Töö käigus selgus, et olem-tunnus-väärtus mudel on parim salvestuste jaoks ja seminormaliseeritud mudel pärimise jaoks. Kõige enam valmistas raskusi seminormaliseeritud mudeli realiseerimine ning programmeerimise liidese tegemine. Ühtlasi oli seminormaliseeritud ka kõige huvitavam uurida, sest selle kohta puudus igasugune avalik materjal ning ainuke infoallikas oli juhendaja Targo Tennisberg.
Summary The goal of this thesis was to analyze the possibilities of implementing a dynamic datastructure and it’s performance when inserting or selecting data. Because most of the applications developed in these days use database and it is not always possible to define the structure of the data, it is necessary to use a dynamic datamodel. In the thesis seminormalised datamodel was compared to entity-attribute-value datamodel. Their nature was analysed, the placement of data in the tables examplified and the speed of inserts and selects compared. The last part of the thesis brought out which of the models is best for which data operations. During the thesis it was concluded that for applications which mostly insert data, the best dynamic datamodel is entity-attribute-value model and for applications which mostly select data, the best one is seminormalized datamodel. Implementing the interface for seminormalized datamodel turned out to be the most difficult, but also the most rewarding, because of the opportunity to familiarize myself with a subject that lacks public information so far.
Summary The goal of this thesis was to analyze the possibilities of implementing a dynamic datastructure and it’s performance when inserting or selecting data. Because most of the applications developed in these days use database and it is not always possible to define the structure of the data, it is necessary to use a dynamic datamodel. In the thesis seminormalised datamodel was compared to entity-attribute-value datamodel. Their nature was analysed, the placement of data in the tables examplified and the speed of inserts and selects compared. The last part of the thesis brought out which of the models is best for which data operations. During the thesis it was concluded that for applications which mostly insert data, the best dynamic datamodel is entity-attribute-value model and for applications which mostly select data, the best one is seminormalized datamodel. Implementing the interface for seminormalized datamodel turned out to be the most difficult, but also the most rewarding, because of the opportunity to familiarize myself with a subject that lacks public information so far.