Semantic Caching for OLAP via LLM-Based Query Canonicalization (Extended Version)

Laurent Bindschaedler · 2026

Read Semantic Caching for OLAP via LLM-Based Query Canonicalization (Extended Version) by Laurent Bindschaedler (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-19811.txt.

Description

Dashboard queries repeat themselves semantically across tools, yet most caches key entries by surface text or syntax tree, so equivalent queries miss each other. A 2026 arXiv paper by Laurent Bindschaedler, working alone, introduces a cautious middle-layer cache for dashboard-style OLAP on star-shaped schemas.

The cache canonicalizes both SQL and natural-language queries into one key space, an OLAP Intent Signature that records measures, grouping levels, filters, and time spans. Reuse demands an exact intent match under strict schema checks and confidence-gated acceptance of natural language. Two correctness-preserving derivations, roll-up and filter-down, widen coverage without approximate matching. Across three benchmarks the cache hits far more often than text or syntax keys, with zero false hits.

Catalog record

CREATORLaurent Bindschaedler
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-19811

Browse the full library · About the Tower of Babel