This thesis continues research exploring the benefits of using functional databases based around the functional data model for advanced database applications-particularly those supporting investigative systems. This is a growing generic application domain covering areas such as criminal and military intelligence, which are characterised by significant data complexity, large data sets and the need for high performance, interactive use. An experimental functional database language was developed to provide the requisite semantic richness. However, heavy use in a practical context has shown that language extensions and implementation improvements are required-especially in the crucial areas of string matching and graph traversal. In addition, an implementation on multiprocessor, parallel architectures is essential to meet the performance needs arising from existing and projected database sizes in the chosen application area.