This is a method for accessing and updating views and other relations of a relational database that: (1) does not require any distinction between base and derived expressions of the data elements; (2) provides data independence; (3) ensures predictable behavior of all updates; and, (4) provides uniformity for all updates independent of data source or type. The method further permits updating of all relations (both base and derived) without unnecessary duplication, overlap, or the creation of null elements, and without using unnecessary physical memory for the resulting data elements.