Data Posting: a New Frontier for Data Exchange in the Big Data Era.
Domenico Saccà, Edoardo Serra · 2013
Data exchange [5, 1] is the problem of migrating a data instance from a source schema to a target schema such that the materialized data on the target schema satisfies a number of given integrity constraints (mainly inclusion and functional dependencies). The target schema typically contains some new attributes that are defined using existentially quantified variables: the main issue is to reduce arbitrariness in selecting such variable values. Therefore a data exchange solution is required to be “universal ” in the sense that homomorphisms exists into every possible solution, i.e., a universal solution enjoys a sort of “minimal arbitrariness ” property. The main research goal of the large data exchange literature is to single out situations for which a universal solution exists and can be computed in polynomial time. Recently a different approach to data exchange has been proposed in [11] that considers a new type of data dependency, called count constraint (an extension of cardinality constraint), that prescribes the result of a given count operation on a relation to be within a certain range. We illustrate this approach by means of an example. Consider a source relation scheme S with the following attributes: I (Item), B (Brand), P (Price).