Efficient UCC discovery for high-dimensional data with large result sets

Improve the performance of the DUD algorithm for high-dimensional data sets with very large unique column combination result sets.

Background

The paper introduces DUD, a disk-resident unique column combination discovery algorithm designed for limited-memory environments. Its performance depends partly on candidate management and validation costs, which can increase substantially when the number of attributes or the number of discovered UCCs is large. The authors specifically identify improving performance in this regime as an unresolved problem, rather than merely a general direction for future work.

References

Improving performance on high-dimensional data sets with very large UCC result sets is also an open problem.

— Efficient discovery of unique column combinations on disk-resident data with limited memory  (2609.00783 - Wan et al., 1 Sep 2026) in Section 5, Discussion, subsection “Open problems”