Schema Exchange: A Template-Based Approach to Data and Metadata Translation

Abstract

Abstract. We study the schema exchange problem, a natural extension of the data exchange problem to an intensional level. To this end, we first introduce the notion of schema template, a tool for the representation of a class of schemas sharing the same structure. We then define the schema exchange notion as the problem of (i) taking a schema that matches a source template, and (ii) generating a new schema for a target template, on the basis of a set of dependencies defined over the two templates. This framework allows the definition, once for all, of generic transformations that work for several schemas. A method for the generation of a "correct" solution of the schema exchange problem is proposed and a number of general results are given. We also show how it is possible to automatically generate a data exchange setting from a schema exchange solution. This allows the definition of queries to migrate data from a source database into the one obtained as a result of a schema exchange.

Migrating Data Semantic from Relational Database System To NOSQL Systems to Improve Data Quality for Big Data Analytics System

“…In the article [4], the authors define the NoSQLayer as a framework whose objective is the migration of RDBMS represented by MySQL to the NoSQL system represented by MongoDB by dividing its processing into two stages: metadata extraction and data mapping through an abstract model. In article [5], the authors translate data between different formats using the principles of WML data and also a combination of XML transformation functions. In article [6], the authors propose an automatic way of transforming the SQL system represented by MySQL to a NoSQL system represented by the column-oriented category and specifically the HBase tool.…”

Schema exchange: Generic mappings for transforming data and metadata

“…Some of the results of this paper have been presented, in a preliminary form, in [26]. In this paper we have extended our earlier work in several ways: we have added a lot of practical examples to make more clear the various notions and results; we have added a new section (Section 2) that illustrates a motivating scenario and provides an overview of the approach; we have refined the notions of schema exchange, encoding and decoding of e-schemas, s-t dependencies and metaroute (now called metalink); we have introduced a new and more general S-D transformation process (now the new algorithm is self-contained as it does not refer to notions and techniques defined elsewhere); we have included the proofs for all the results; we have added a new section (Section 7) discussing related works.…”

Data exchange with data-metadata translations

“…Lastly, the problem of generating a target schema using mappings from a source schema (i.e., metadata-to-metadata translations) is also known as schema translation [17]. The ModelGen operator of Model Management [2] (see Section 3.2 of [1] for a survey of ModelGen work) relies on a library of pre-defined transformations (or rules) that convert the source schema into a target schema.…”

“…The ModelGen operator of Model Management [2] (see Section 3.2 of [1] for a survey of ModelGen work) relies on a library of pre-defined transformations (or rules) that convert the source schema into a target schema. Alternatively, [17] uses mappings between meta-metamodels to transform metadata. None of these approaches, however, allow for data-to-metadata transformations.…”