A recent arXiv preprint (2610.22705) addresses the challenge of identifying similar queries at runtime, a task that underpins workload analysis, detection of recurring behavior, auditing, debugging, and anomaly detection. The authors propose an approach based on data provenance aggregation and structural descriptors, aimed at query-template identification.

The abstract briefly contrasts this with existing methods for evaluating query similarity, but the source text is truncated at that point. As a result, the specific advantages or experimental outcomes of the proposed method are not available from the provided material. The paper is listed as a cross-type announcement, though no further publication details are given.

Given the limited information, this summary can only reflect the stated motivation and the proposed combination of techniques. Readers interested in the full methodology would need to consult the complete paper on arXiv.