A Generalized DBMS to Support Diversified Data.
Abstract
This project supported an investigation into extending a relational system to support new kinds of data objects such as lines, points, polygons, bit maps, text, documents, vectors and arrays. Two important tactics were studied: 1) use of an abstract data type system (ADT); and 2) use of procedural objects. Basically an ADT system allows new data types to be added to a relational system to augment the normal character strings, integers and floating point numbers traditionally provided. In addition, procedures provide a powerful data modeling capability that is especially useful for complex objects or objects that have unpredictable composition. The important results that the authors achieved are twofold. First, they obtained results on the integration of an ADT system into a general purpose DBMS. In particular, they discovered how to integrate new data types with query optimization routines and how to allow new access methods to be constructed for the new data types. Second, they designed a syntax for efficient manipulation of procedural objects and then constructed query optimization algorithms to efficiently process this extended syntax.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 21, 1987
- Accession Number
- ADA188111
Entities
People
- Eugene C. Wong
- Lawrence Rowe
- Michael Stonebraker
Organizations
- University of California, Berkeley