A Database Array Algebra for Spatio-Temporal Data and Beyond

  • Authors:

  • Peter Baumann

  • Affiliations:

  • -

  • Venue:
  • NGIT '99 Proceedings of the 4th International Workshop on Next Generation Information Technologies and Systems
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Recently multidimensional arrays have received considerable attention among the database community, applications ranging from GIS to OLAP. Work on the formalization of arrays frequently focuses on mapping sparse arrays to ROLAP schemata. Database modeling of further array types, such as image data, is done differently and with less rigid methods. A unifying formal framework for general array handling of image, sensor, statistics, and OLAP data is missing. We present a cross-dimensional and application-independent algebra for the high-level treatment of arbitrary arrays. An array constructor, a generalized aggregate, plus a multidimensional sorter allow to declaratively manipulate arrays. This algebra forms the conceptual basis of a domain-independent array DBMS, RasDaMan, which offers an SQL-based query language with extensive algebraic query and storage optimization. The system is in practical use in neuro science. We introduce the algebra and show how the operators transform to the array query language. The universality of our approach is demonstrated by a number of examples from imaging, statistics, and OLAP.