The composition of DaaS (Data-as-a-Service) services is a powerful solution for building value-added applications on top of existing ones. However, privacy concerns are still among the key challenges that keep hampering DaaS composition. Indeed services may follow different, conflicting privacy specifications with respect to the data they use and provide. In this paper, we propose an approach for privacyaware composition of DaaS services. Our approach allows specifying privacy requirements and policies and verifying the compatibility of services involved in a composition. We propose an adaptation protocol that makes it possible to reconcile the privacy specifications of services when incompatibilities arise in a composition. We validate the applicability of our proposal through a set of experiments.