Abstract. We consider the problem of designing data wrapping ontologies whose purpose is to describe relational data sources and to provide a semantically enriched access to the underlying data. Since such ontologies must be close to the data they wrap, the new terms that they introduce must be “supported” by data from the relational sources; i.e. when queried, they should return nonempty answers. In order to ensure nonemptiness, those wrapping ontologies are usually carefully handcrafted by taking into account the query answering mechanism. In this paper we address the problem of supporting an ontology engineer in this task. We provide an algorithm for verifying emptiness of a term in the data wrapping ontology w.r.t. the data sources. We also show how this algorithm can be used to guide the ontology engineer in fixing potential terms unsupported by the data. Finally, we present an implemented tool and an empirical study showing benefits of our approach.