Treating hierarchical data structures in the SDC Time-Shared Data Management System (TDMS)

Abstract

An important consideration in the design of programming systems for the management of large files of data is the method of treating hierarchical data (that is, data among which logical relationships exist at more than two levels). Recent systems have accomplished this by simply duplicating some essential item of data at several levels. Such duplication makes the storage of even small data bases inefficient; for large masses of data, storage becomes economically unfeasible. Other systems have provided the means to specify and construct hierarchies, but none have provided language that affords control over retrieval and output levels, and control over the scope of output. Within the Time-Shared Data Management System (TDMS), currently being produced at System Development Corporation, a method has been devised for maintaining hierarchical associations within logical entries of a data base. Basically, the technique permits the automatic association of related data through a device known as a repeating group. The term “repeating group” is not new, but the TDMS treatment of the repeating group concept is. This paper describes how this technique is implemented in the language and tables of TDMS.