Statistical databases

Most data collected in statistics and science is still stored in simple flat files, usually data matrices with rows identified by a case identifier (case_id), columns corresponding to attributes (variables), and numerical data types for the elements of each matrix due to coding of all attributes involved. Each row (tuple) carries the (coded) values of the attributes, besides the case_id. Due to a suitable encoding that maps a natural domain to numerical ones, all matrix entries have a numeric data type. The scales of the attributes may of course be quite different.