ID.10. DBMS Augmentation Using the Web of Data

The Web consists of a huge number of HTML tables containing relational-style data and schema-like descriptions such as labelled and typed columns. Although, these web tables are a valuable resource of easily accessible structured information, state-of-the-art database systems do not provide support to integrate, reuse and analyze this data in an ad-hoc manner. To overcome this problem we built an extended DBMS called DrillBeyond (see demo here) that introduces entity augmentation as a novel query type, enabling users to query a local database together with a large web table corpus in a seamless manner. This unburdens the users from tasks such as finding the best open web tables that fit their needs, downloading and integrating them into the local database schema and rewriting their own queries. Whereas the entity augmentation operation already allows extending local tables with additional attributes on the fly we would like to push the idea of an open and augmentable DBMS even further. Here we can think about more complex horizontal augmentations that, given a local table, add all related attributes and the corresponding data published on the Web until a certain density threshold is reached. Further we want to investigate vertical augmentations (in contrast to the horizontal case) that allow users to define just the table header and optionally some example tuples and the DBMS than automatically fills the table with tuples coming from the Web. On the other side, augmentation could also support the users in defining DDL statements where the attribute names are automatically inferred from existing schema definitions. To evaluate the concepts invented in this topic we could rely on our already existing Dresden Web table Corpus consisting of 112M tables.

Main Advisor at Technische Universit├Ąt Dresden (TUD)
Co-advisor at Universitat Polit├Ęcnica de Catalunya (UPC)