Introduction: The Magic of Constraint Databases
Rectangle(x, y) :- x >= 0, x <= 10, y >= 0, y <= 20.
Each constraint tuple is a conjunction of inequalities using the same set of variables, in the rectangle example the spatial variables x and y. It is understood to mean that any (x,y) point that satisfies the inequalities is in the rectangle. There are an infinite number of such points. A constraint relation is a finite set of constraint tuples. A constraint database is a finite set of constraint relations. Kanellakis, Kuper, and Revesz introduced constraint query languges that accept constraint databases as input and give constraint databases as output. This constraint data model is a natural generalization of the relational data model to finitely representable infinite relations. Of course, each constraint tuple is finitely representable.
Should those who write queries worry about the representation of the infinite relations? NO, use schema of the (infinite) relations like in relational databases. Is it more difficult to write SQL queries for constraint databases than for relational databases? NO Can one do more with SQL and constraint databases than with SQL and relational databases? YES
We can do more with constraint databases using the same effort of writing SQL queries. It seems magic, but this is not. It comes from the power of infinity of the relations. Learn about it more from the references below. |
|
Bibliography
|
Mailing List
· Subscribe and unsubscribe mailing list at: |
Conferences on Constraint Databases |
Systems & Applications |
Books
|
|