schema-mapping - heterogenity operations

To find a common format to conversion of local fields into a common global fields value is one side. Yet there might be semantic heterogenity in that some enumeration is not given as fields values but structure separations. To get a decent mapping of these we need to instantiate structure information as a table and allow them to combine on queries. And there we need an engine that maps the enumerated values into the different channels to different tables and databases.

This specific area of schema mapping has been investigated quite a number of times. In a datalog query it means that a relation name might be used as an argument name as well. In other words

 A(C,D) & X(A,D) 

 A(C,D) -> A(C,D) & tables(A) 

The most well-known prior art in this area is SchemaSQL which extends common SQL script with additional syntax to declare variables that range over columns of a table or simply over all tables reachable from a query. Some later work (as noted above) was rewriting the instantions of these meta-variables with first-order queries to a table representing the structure of the tables and databases involved. Such a structure table could be built on the fly from retroinspection of wrappers.

And last not least we should remember optimizations in this area since the execution of a query with meta-variables might be seen to take the union over all involved tables in the range of the meta-variable - or just cut it down to the ones being narrowed by a later "where"-clause. That makes for a much smaller series of queries to the different tables implicitly attached through the meta-variable.

Operations

• schema structure discovery - wrapper retroinspection
• meta-variable expansion - materialized view of structure
• meta-variable subsitution - rule rewriting
• query subdivion - 1-to-n queries and aggregations