Lina Al-Jadir

Once Upon a Time a DTD Evolved into Another DTD ...

XML has become an emerging standard for data representation and data exchange over the web. In many applications a schema is associated with an XML document to specify and enforce the structure of the document. The schema may change over time to reflect a change in the real-world, a change in the user’s requirements, mistakes or missing information in the initial design. In this paper, we consider DTDs as XML schema mechanism, and present an approach to manage DTD evolution. We build a set of DTD changes. We identify invariants which must be preserved across DTD changes. We define the semantics of each DTD change such that the new DTD is valid, existing documents conform to the new DTD, and data is not lost if possible. We illustrate our approach with a scenario.

F2/XML: Storing XML Documents in Object Databases

In this paper we propose a new method to store an XML document in an object database (DB). First, the documents DTD is mapped into a DB schema, then the XML document is mapped into a DB instance. Our method stores the element-subelement relationships and the element attributes. It takes into account the order and the cardinality of subelements. It keeps track of the groupings in an element, and alternatives among subelements. Our method allows us also to retrieve back the DTD and the entire document from the database without loss of information. We implemented our method in the F2 database system, and tested it on sample XML documents.

Transposed Storage of an Object Database to Reduce the Cost of Schema Changes

Modifying the schema of a populated database is an expensive operation. We propose to use the non-classical transposed storage of an object database. The transposed storage avoids database reorganization and reduces the number of input/output operations in the context of schema evolution. Thus schema changes are not anymore costly operations. Consequendy immediate and physical propagation of schema changes can be supported. We extend the 007 benchmark with schema evolution operations and submit our F2 DBMS to this benchmark. The obtained results demonstrate the feasibility and performance of our approach.

Encapsulating classification in an OODBMS for data mining applications

Classification is an important task in data mining. Encapsulating classification in an object-oriented database system requires additional features: we propose multiobjects and schema evolution. Our approach allows us to store classification functions, and to store instances of each group in order to retrieve them later. Since the database is operational, it allows us also to perform dynamic classification, i.e. add/remove instances to/from groups over time. Moreover it allows us to update classification functions (if we choose another population sample or apply another classifier) and have the instances of groups consequently reclassified. We illustrate our approach with a target mailing application.

Integrating Association Rule Mining Algorithms with the F2 OODBMS

In this paper, we integrate two association rule mining algorithms, Apriori and TRAND, with the F2 object-oriented database system (DBMS). The advantages of our integration are the following. Both algorithms do not need to maintain complicated data structures and use only database classes. Both algorithms do not need to manage the buffer since it is handled by the DBMS. Both algorithms store frequent itemsets in the database which can be retrieved later using the DBMS data manipulation language. In addition to that, the TRAND algorithm takes advantage of the transposed storage supported in F2. To compute the support of candidate itemsets, it applies logical AND on boolean attributes, implemented in F2 as vectors, and avoids scanning the database. This reduces significantly the number of block accesses and consequently the execution time.

Implementing NRDR Using OO Database Management System (OODBMS)

Motives for using a database management system (DBMS) to build a knowledge base system (KBS), include KBSs lack of ability to manage large sets of rules, to control concurrent access and to manage multiple knowledge bases simultaneously. In work, we build a KBS using a database management system DBMS for its schema evolution ability. We use this ability of an Object Oriented DBMS (OODBMS) to manage the consistency of an incrementally built knowledge base (KB).The significance of this work is two folds: first, it provides an efficient mechanism maintaining consistency of an evolving classification hierarchy, using built-in schema evolution features of an OODBMS. Second, it enhances the interface of an OODBMS, to allow intelligent classification queries over stored objects.

F2/XML: Navigating through Linked XML Documents

As eXtensible Markup Language (XML) has become anemerging standard for data representation and data exchange on the Web, there is an increasing trend to store XML data in databases, in order to use the mature technology of database systems. The approaches which have been proposed so far handle non-linked XML documents.One of the reasons for the phenomenal growth of the Internet is the ability for documents to link to other documents. Therefore, linkingcapabilities for XML documents have recently been proposed using XLink and XPointer. In this paper, we present the F2/XML method tostore linked XML documents in object databases and navigate through them. We illustrate our method with an example.