Optimizing subset queries: a step towards SQL-based inductive databases for itemsets