Instance-Optimized String Fingerprints

TL;DR

This paper introduces optimized string fingerprints as a lightweight indexing method to improve string query processing in cloud data warehouses, achieving significant speedups in table scans.

Contribution

It presents a novel approach to optimize string fingerprints for specific workloads using mixed-integer optimization, enhancing their effectiveness and generalization.

Findings

Up to 1.36× speedup in table scans on IMDb data

Optimized fingerprints reduce compute and I/O overhead

Method generalizes to unseen table predicates

Abstract

Recent research found that cloud data warehouses are text-heavy. However, their capabilities for efficiently processing string columns remain limited, relying primarily on techniques like dictionary encoding and prefix-based partition pruning. In recent work, we introduced string fingerprints - a lightweight secondary index structure designed to approximate LIKE predicates, albeit with false positives. This approach is particularly compelling for columnar query engines, where fingerprints can help reduce both compute and I/O overhead. We show that string fingerprints can be optimized for specific workloads using mixed-integer optimization, and that they can generalize to unseen table predicates. On an IMDb column evaluated in DuckDB v1.3, this yields table-scan speedups of up to 1.36$\times$.