relational-contingency-table

Computes contingency tables for relational databases, i.e. counts across tables

Input

• A relational database data-db
• A metadata database setup-db
• An output database ct-db
• A list of first-order terms (aka functor nodes, first-order random variables) stored in the table setup-db.FunctorSet
• Optional: A list of first-order variables stored in the table setup-db.FunctorSet

Output

A table ct-db.ct-table of the form (fill in table format exaple) where

• A row lists a combination of possible values for each first-order term
• Column count indicates the number of database instances where each first-order term is assigned the value in the row. Here count is obtained from the size of the result set of the database query defined by the conjunction of literals where the term = value.
• If first-order variables are specified, the contingency tables gives counts for each grounding or instantiation of the first-order variables with constants denoting individuals.

References

The algorithm implements the method from (Qian et al. CIKM 2014) based on metaqueries and the Inverse Fast Moebius Transform.