Database Paper Browser

Back to papers

Can Datalog be approximated?

Summary: Introduces a quantitative error notion and upper/lower-envelope constraints to study approximating recursive Datalog predicates by finite unions of CQs, contrasting absolute vs relative approximation. Proves absolute approximation impossible for unbounded Datalog, rules out constant-relative-error CQ approximations for broad unbounded classes, and gives sharp FO inapproximability for transitive closure and cycle queries. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID
1020
Venue
PODS
Year
1994
Pagerank
4.9352222e-05
Overall Rank
6,751 | 53.09%
DOI
-

Incoming Non-self Citations Over Time

Authors

Incoming Citations (Sorted by Pagerank)

Showing 4 of 4 citing papers.

Rank Citing Paper Year Venue Pagerank
4,405 Filtering with Approximate Predicates 1998 VLDB 6.2048285e-05
6,312 Type Inference for Datalog and its Application to Query Optimisation 2008 PODS 5.1109656e-05
12,704 Approximate Query Translation Across Heterogeneous Information Sources 2000 VLDB 4.1905499e-05
12,801 Combinatorial Games in Database Theory 1995 PODS 4.1905499e-05
Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 2 of 2 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank Cited Paper Year Venue Pagerank
1,532 The Parallel Complexity of Simple Chain Queries (Extended Abstract) 1987 PODS 0.00011466849
3,889 On the Expressive Power of Datalog: Tools and a Case Study 1990 PODS 6.6570358e-05
Previous Page 1 / 1 Next

Semantically Similar Papers