Re: OT look at my project AlgebraDB :)

Peter Hanusiak wrote:
>
> Our tables are classic tables which haven't any problem, infinite
> tables are built in function like add which have 3 attribute x,y,z
> and this table corespond to predicate x+y=z. so you must deliver to
> this table at least 2 atributes. for example if you ask for match for
> x=5 y=4 z=10 you get answer false. and next kind of table is defined
> table, something like equation solving for example we can define new
> table "table"
>
> define table{
> temp(x,y) project(a,b,x)*add(x,y,z)*result(c,y,z)
> table(a,x) temp(x,y)*emp(x,a)
> }

Am I right in assuming that this is basically the same as the following datalog?

 temp(x,y) <- project(a,b,x), add(x,y,z), result(c,y,z).  table(a,x) <- temp(x,y), emp(x,a).

And what about query optimization? There is quite some literature on doing that for datalog.

> Yes but we must have safe formulas.

Is that a problem?

-- 
  Jan Hidders