The volume and variety of modern day databases presents a particular challenge to the system testing community. The question is how to go about testing such large collections of various data types ranging from tables to texts and images. To test those applications which use them, these conglomerations of multiple data object types have to be automatically generated and validated. There is no other way but to automate the test process. This contribution outlines the challenge and presents an automated approach to setting up and testing big data bases. At the end a case study of a large data warehouse is discussed with lessons learned from that industrial test project.