Accessing web forms from an object- relational database system

A main memory object-relational database system, AMOS II, has been developed at Uppsala Database Laboratory (UDBL). The system provides common database facilities and a powerful query language but also, through it’s mediator-wrapper approach, features for the combination of data from heterogeneous data sources. The AMOS II query processor is extensible through a generalized foreign function mechanism. Currently AMOS II has interfaces to the programming languages C, Lisp and Java implying that applications working against AMOS II can be developed in one of these programming languages. Unfortunately, accessing information from the web cannot be done in the same way as using conventional databases. For the extraction of data from the web there exists Internet wrapper toolkits to specify programmatic interfaces to more or less well structured web sources. The thesis is about developing an extension to the existing AMOS II system, ORWIF (Object Relational Wrapper of Internet Forms), that facilitates the combined access to data from the web with data from other data sources e.g databases. In the ORWIF project data is extracted from a web form using foreign functions, existing Java libraries and a publicly available Internet wrapper toolkit. A description and comparison among three existing Internet wrapper toolkits is also included. My solution offers a flexible and easy way for users to access and analyse information retrieved from web forms and combining it with data from other sources through an OO mediator database system. For example, users can do their own “price running” with ORWIF. In my work I addressed the issue of optional input fields in web forms and how this can complicate extracting information from web sources. The problem was solved through delegating the responsibility of handling these, so called, omitted parameter values to the code of the foreign function.