Open Access

Fabian Panse, Meike Klettke, Johannes Schildgen, Wolfram Wingerath

• A flexible and versed generation of test data is an important aspect in benchmarking algorithms for data integration. This includes the generation of heterogeneous schemas, each representing another data source of the integration benchmark. In this paper, we present our ongoing research on a novel approach for similarity-driven generation of schemas, which takes as input an arbitrary dataset, extracts its schema, and derives a set of output schemas from it. In contrast to previous solutions, we do not focus on structural transformations of relational or XML schemas, but extend the scope to contextual transformations and NoSQL data models, where the required schema information is often only implicitly defined within the data and must first be extracted. In addition, we utilize a novel method that generates multiple schemas based on user-defined heterogeneity constraints making the generation process configurable even for non-experts.