Guardando históricos de dimensões em data warehouses.

Abstract

One of the most important patrimony of an organization is its information. This patrimony is nearly always kept in the form of operational systems, or production systems. Broadly speaking, a production system works on updated and detailed data. Its main objective is to give support to the operations of the organization in the everyday life, or to its managing in the operational level. The higher levels of management in the organization need summarized and historical information, and are not, thus, covered by production systems. The incapacity of production systems in supporting managing activities on an organization in all its levels has forced intense research and development efforts towards building decision supports systems stricto sensu, called Data Warehouses. In spite of the huge progress in this direction, many of the requirements of these systems sill lack adequate solutions. In this work, we deal with the requirement time in Data Warehouses, presenting a formal solution for the problem of the dimensions attributes history in Star Schemas. The motivation for our work came from the awareness that, although the databases for decision support systems should be temporal by definition, the treatment that has been given to time (data structures and temporal interfaces) is still rudimentary.