MutShrink: um método de redução de banco de dados de teste baseado em mutação
Nenhuma Miniatura disponível
Data
2017-08-11
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Goiás
Resumo
Regression testing for database applications can be a computationally costly task as it
often deals with databases with large volumes of data and complex SQL statements (for
example, nested queries, set comparisons, use of functions and operators). In this context,
some works only select a subset of the database for testing purposes, that is, select data
to create a test database and thus improve test efficiency. But usually, the selection of
test data is also a complex optimization problem. Thus, this work proposes a method of
selecting test data for regression testing on SQL statements based on mutation analysis,
called MutShrink. The goal is to minimize the cost of testing by reducing the size of the
database while maintaining the same effectiveness as the original database. MutShrink
consists of using the result of the generated mutants to evaluate the database and select
tuples using filters in these results, selecting reduced sets of test data. Experiments
were performed using a benchmark with complex SQLs and database with large data
volume. We compared our proposal with the QAShrink tool and the results revealed
that MutShrink overcame the QAShrink tool in 92.85 % of cases when evaluated by the
Mutation Score metric and 57.14 % of cases when evaluated by the metric Full Predicate
Coverage.
Descrição
Citação
TOLEDO, L. I. MutShrink: um método de redução de banco de dados de teste baseado em mutação. 2017. 66 f.
Dissertação (Mestrado em Ciência da Computação) - Universidade Federal de Goiás, Goiânia, 2017.