No abstract

Abstract. This paper presents a new framework for users to select relevant data from an XML document and store it in an existing relational database, as opposed to previous approaches that shred the entire XML document into a newly created database of a newly designed schema. The framework is based on a notion of XML2DB mappings. An XML2DB mapping extends a (possibly recursive) DTD by associating element types with semantic attributes and rules. It extracts either part or all of the data from an XML document, and generates SQL updates to increment an existing database using the XML data. We also provide an efficient technique to evaluate XML2DB mappings in parallel with SAX parsing. These yield a systematic method to store XML data selectively in an existing database.

Over a traditional Database Management System (DBMS), the answer to an aggregate query is usually much smaller than the answer to a similar non-aggregate query. Therefore, we call such a query "condensative". Current proposals for declarative query languages over data streams do not support such condensative querying.In order to make existing stream query languages more expressive so that they enable a user both, to state more intuitively interesting queries, and to support condensative querying, we propose a new data stream model, referred to as the sequence model, and an extension to SQL-like query languages by operators that allows one to specify the frequency by which a query returns answer tuples. We show that such frequency operators allow one to express sampling over streams. If combined with existing sliding window operators, they support queries with "jumping windows".We show with a number of examples from a sensor monitoring application how complex queries can be elegantly formulated in a stream query language with frequency operators.

No abstract