Um estudo comparativo sobre uso de modelos de dados para notas fiscais eletrônicas

Abstract

The term Big Data has become increasingly popular, and consists of a large volume of data and, therefore, needs special attention, such as the use of unconventional tools, which have the ability to store, process and integrate all information of a huge database. It is common these days, companies and organizations to adopt these technologies, as we currently live in the information age and the volume of data grows exponentially. This also occurs in the accounting world, especially from 2007, when the Public System of Digital Bookkeeping was implemented in Brazil, through the Federal Government, in private sector companies and also in public government institutions, responsible for the inspection and control of these activities such as municipal and state tax authorities and the Federal Revenue Service. This has led to the constant use of a large volume of data with electronic invoices among these agencies, making it necessary for them to have a structured and robust data model for the storage, processing and use through queries of these electronic fiscal documents. This study analyzed the performance of two data models: relational and document-oriented, using PostgreSQL and MongoDB respectively, executed by a query benchmark, using a batch of 50 and 100 thousand electronic invoices (NFes) and having up as a response time metric. It was obtained as a result that MongoDB has a higher efficiency rate in the query response time than PostgreSQL, regardless of the analyzed operator. This research brought two main contributions: the possibility of measuring the performance of two data models, using a database of NFes in the context of Big Data, as well as establishing, through software, a better understanding of the fields of an NF- and.