Data Provenance Aggregation and Structural Descriptors for Query-Template Identification
Runtime query similarity identification supports several activities, including workload analysis, identifying recurring behaviour, auditing, debugging, and anomaly detection.
ProofPaper ↗
Key points
- Existing methods for evaluating query similarity mostly rely on SQL text, execution plans, or result values but do not account for input data or how results are derived during query execution.
- This paper proposes a novel approach using data provenance for query-template identification.
- We construct query-level profiles that capture explicit data dependencies, occurrence frequencies, and structural properties by aggregating tuple-level Lineage and Why-provenance across entire query results.
- We evaluate our proposals using these profiles to identify query templates over \(12{,}500\) distinct query instances generated from parametrised TPC-DS query templates.
Sources (1)
- [1]Data Provenance Aggregation and Structural Descriptors for Query-Template IdentificationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 10:20 PM
Runtime query similarity identification supports several activities, including workload analysis, identifying recurring behaviour, auditing, debugging, and anomaly detection.
Existing methods for evaluating query similarity mostly rely on SQL text, execution plans, or result values but do not account for input data or how results are derived during query execution.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 15, 2026SAG2: Embedding-Free Multi-Hop Retrieval with BM25 and JEV
- Oct 15, 2026EliteReplay: Preserving Diverse Solutions in RL Post-Training via Quality-Diversity Search
- Oct 15, 2026PremiseSQL: Evaluating Premise Critique and Abstention in Text-to-SQL Systems
- Oct 15, 2026Each Reward to Its Relevant Tokens: Reward-Specific Token-Level Credit Assignment for Multi-Reward RL
- Oct 15, 2026RDL-Bench: A Benchmark for Task-Aware Question Answering over Relational Databases
- Oct 14, 2026mistralai/Timestral-209M