Value-based queries

• RDBMSs are well-suited for performing queries on the values of attributes

  • efficient and suited to parallel processing

• ODBMS vendors are evolving techniques for optimising such queries

  • clustering techniques
  • specialized indexing mechanisms
  • local caches

Previous slide

Next slide

Back to first slide

View graphic version

OO GIS support highly efficient spatial indexing structures. But as with earlier-generation hierarchical and network databases, ODBMS' biggest performance hit is in processing queries based on the values of objects' attributes. This is not a fundamental problem - the flexibility of the more modern OODBMSs allows an index to be set up on any attribute independent of any other: the index is then an object like any other and can be distributed to local client caches in the normal way.

Some other ways OO-GIS vendors are improving the performance of queries and sorting include:
• clustering an object's components together on the disk
• supporting efficient join indexing techniques originally developed for RDBMSs

in addition to:
• maintaining additional non-spatial indexing structures
  
  

4/98