An extended relational algebra with control over duplicate elimination

Abstract

In the pure relational model, duplicate tuples are automatically eliminated. Some real world languages such as DAPLEX, however, give users control over duplicate elimination. This paper extends the relational model to include multiset relations, i.e., relations with duplicate tuples. It considers three formalisms for expressing queries in this model: extended relational algebra, tableaux, and DAPLEX. It shows that, as in the original algebra, the equivalence problem for conjunctive expressions in the extended algebra can be solved using tableaux, and is NP-complete. Finally, it demonstrates that the extended algebra and DAPLEX have essentially the same expressiveness relative to conjunctive expressions.