Angra-db : indexing and query language

Abstract

This work presents the process of development of a query language into a NoSQL documental database called Angra-DB. Angra-DB is an open source project that is developed at University of Brasilia and has requirements about its features extensibility, modularity and ease of use. The steps required in order to enable the query language building according to the project’s principles involved the creation of an indexing feature for the database and the establishment of a common querying API, which would be used by query languages in order to perform queries into the stored documents. At the end of this implementation a query language inspired in JSON Query Language’s specifications was built, allowing rich queries to be performed into the database. The testing of the proposed architeture showed that, while its performance during insertions of new documents lacked in comparison with other document-oriented databases such as CouchDB, the execution times of the queries obtained similar results. Moreover, the query language implementation focused more in its readability and ease of use, when compared to other query languages used in documental databases. Future works involves the further improvement of the architecture’s performance, the improvement of the querying capabilities and the creation of new query languages, possibly to specific data domains.