by Ming X
The article I read is called the development of ordered SQL packages to support data warehousing, by Wilfred Ng and Mark Levene. “Data warehousing is a corporate strategy that needs to integrate information from several sources of separately developed Database Management Systems.” The authors mention that future database management systems should provide adequate facilities to manage a wide range of information arising from such integration. Since the order of data is usually involved in business queries; users can extend the relational model to incorporate partial orderings into data domains and describe the ordered relational model. User can use OSQL, which allows querying over ordered relational databases. Ordered SQL (OSQL) is an extension of the Data Definition Language (DDL) and Data Manipulation Language (DML) of SQL for the ordered relational model. OSQL can be applied to solve various problems that arise in relational DBMSs involving applications of temporal information, incomplete information and fuzzy information under the unifying framework of the ordered relational model. There are three OSQL packages: OSQL_TIME, OSQL_INCOMP and OSQL_FUZZY.
Data warehousing are built upon a client-server architecture, the programming approach has to pay the performance penalty in the data extraction process,if there are too many calls from the programming level to the ralational level.
OSQL offers the advantage that can help to relieve the burden of the bandwidth of a network system and the loads of client processes. It allows us to modularize a collection of generic operations on an ordered data domain. These operations can then be called from within OSQL whenever the package they belong to is loaded into system.
Ng, Wilfred and Levene, Mark: The development of ordered SQL packages to support data warehousing, Journal of Database Management 12. 4 (Oct-Dec 2001): 26-39.