Complexity of Deciding Query Relevance in Databases
A recent study published on arXiv (2604.22422) examines the challenges associated with assessing the relevance of a fact to a Boolean conjunctive query (CQ) in a database. The central question is whether a specific fact is part of the smallest subset of the database that fulfills the query’s requirements. The research reveals that relevance is Σ^p_2-complete for CQs utilizing a binary signature, indicating it is more complex than query evaluation. NP-hardness is already established for acyclic chain CQs. The authors highlight that self-joins (where multiple atoms share the same relation) significantly contribute to this complexity: when self-joins are restricted or limited, the complexity of relevance aligns with query evaluation (NP without structural constraints).
Key facts
- The problem is deciding if a fact f is relevant to a query q wrt database D.
- Relevance is Σ^p_2-complete for CQs over a binary signature.
- NP-hardness applies already to acyclic chain CQs.
- Self-joins are identified as the cause of increased complexity.
- Forbidding or bounding self-joins reduces relevance complexity to that of query evaluation.
- The paper is published on arXiv with ID 2604.22422.
- The announcement type is cross.
- The study addresses combined complexity.
Entities
Institutions
- arXiv