Metabolomics is defined as the comprehensive analysis of all the metabolites of a biological system, although there is considerable confusion in terminology. The task of metabolomics is not attainable with the current state of analytical science, and various approaches are used to accommodate this limitation and the associated complexity of the metabolome. Major advances and a totally new approach to analysis will be essential. Meanwhile, a number of techniques, namely, metabolite profiling, metabolite fingerprinting and target analysis are combined to make this task possible. There is an implicit assumption in these approaches that the data are complementary and that the metabolome can be assessed as a summation of individual data. The integrity of metabolomics data depends on state-of-the-art sample handling and measurement techniques. These ideas are developed in this review.