Delo z grafi v relacijskih in sodobnih nerelacijskih podatkovnih bazah

MALEJ, BLAŽ (author), Kukar, Matjaž (mentor)

V diplomskem delu smo se lotili raziskovanja grafnih baz, vse skupaj pa smo popestrili s primerjavo med grafnimi in relacijskimi podatkovnimi bazami. Grafne podatkovne baze so baze, ki so primarno namenjene za hranjenje podatkov o grafih in obdelavo le-teh. Graf (kot podatkovna struktura v računalniški znanosti) je množica vozlišč, ki so med seboj povezana s povezavami. Z grafi lahko učinkovito modeliramo različne primere iz sveta okoli nas, npr. socialno omrežje na Facebooku, pri čemer so osebe vozlišča, povezave pa prijateljstva med njimi.\ Odločili smo se, da bomo delali z odprtokodnimi programi, za to smo naredili primerjavo med dvema trenutno najbolj popularnima bazama: grafno bazo Neo4j in relacijsko MySQL. Tema primerjav je bila hitrost izvajanja poizvedb, za primer poizvedbe pa smo si izbrali tipičen problem, ki se imenuje prijatelj prijatelja (angl. textit{friend of a friend}). Ta nastopi, ko hočemo poiskati vse prijatelje osebe v nekem poljubnem socialnem grafu.\ Odločili smo se, da raziskovalni del diplomskega dela popestrimo z dodatnim grafnim SUPB. V našo primerjavo smo vključili še SUPB OrientDB, ki se pogosto pojavlja v primerjavah z Neo4j.

• Language: Slovenian
• Keywords: Grafna podatkovna baza, Neo4j, MySQL, OrientDB
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2016
• PID: 20.500.12556/RUL-81061
• Publication date in RUL: 25.03.2016

Secondary language

• Language: English
• Title: Working with graphs in relational and non-relational databases

Abstract

The thema of this diploma thesis are graph databases and the comparison between graph databases and relational databases. Graph databases are used to store and process all kinds of graph data. Graphs, as a data structure in computer science, are a set of vertices which are connected with things called edges. Using this principle, we can effectively model various real life examples, e.g. social network of Facebook, for which we could store in graph entities of friends as nodes, and friendships as edges.\ We decided to compare two currently most popular open source databases, Neo4j - representing graph databases and relational database named MySQL. We compared the execution time of typical query where one would want to get all the friends of a friend in arbitrary social network (also called friend of a friend problem). In our research we another graph database named OrientDB, which is often used in various comparisons together with Neo4j.

• Keywords: Graph database, Neo4j, MySQL, OrientDB