Artemis - An Automatic Test Suite Generator for Large Scale OLAP Database.

We design an automatic test suite generation tool Artemis for functionality test of Online Analytical Processing Databases (OLAP DBs). This is the first work which accomplishes the work of DB test by integrating three artifacts, i.e., data generation , workload generation and oracle generation , but promises the scalability, effectiveness and efficiency. The key idea of our approach is to design a deterministic random data generation mechanism, based on which we can instantiate the parameterized queries and calculate the oracles simultaneously by resolving the constraint chains along query trees. Since we provide deterministic random functions for data generations corresponding to a predefined schema, repetitive test and data migration become a trivial job. Random workload generation and automatic oracle calculation instead of differential comparison make abundant and massive scale of test possible. We finally provide extensive experiments to show the performance of Artemis .