Resumo:
Relational Databases are the most used models in several applications in
reason of the ease of use in its language of consultation and use in environments
multi-users. With the great volume of information that we have today and, being
that these are increasingly related, databases oriented
graphs as a way to deal with this new demand, given the difficulties of the model
relational to this new scenario. In view of this, this research dealt with the mapping processes,
conversion and migration from the relational model to the graph-oriented model,
above all, the semantic overload of constructors between the two models. The purpose of this
study was the development of an application, called ThrusterDB, that performs this
conversion process from the relational model to the graph-oriented one automatically.
The research contributes by integrating the mapping, conversion and migration phases
from a relational database to a graph-oriented one. This dissertation
presents results that show that the generated database, after the process, provides
a better performance in the average time of consultations carried out, in addition to preserving the semantics
from the source relational database, without any loss or redundancy of
Dice.