next up previous contents
Next: Technological Alternatives Up: Preliminary Conclusions Previous: Multidatabase Query Processing

Assumptions and Problems

Multidatabase querying is based on the assumption that each database in the multidatabase system provides access for ad hoc queries. In addition, constructing a multidatabase system requires detailed documentation of the schemas and semantics of each component database, including information on links to other databases and semantic heterogeneity problems.

A variety of problems were encountered in performing multidatabase queries with OPM*QS, including semantic heterogeneity problems that had to be resolved. For example accession numbers are represented in different formats in different databases: a GDB accession number ``GDB:118746'' is represented as ``G00-118-746'' in GSDB 2.0, while a GSDB accession number ``L05367'' is represented in GDB as ``GSDB:L05367''. OPM*QL supports some general string manipulation and comparison operators in order to resolve such problems. It is hoped that such incompatibilities will normally be hidden in the code for inter-database links, and will therefore not concern users performing multidatabase queries. However merely identifying such incompatibilities, and determining how to resolve them, is a difficult problem, in part because of the lack of detailed documentation for individual database schemas.



& Markowitz
Wed Jan 17 16:39:09 PST 1996