Bootstrapping Learned Cost Models with Synthetic SQL Queries
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914009086689280 |
|---|---|
| author | Nidd, Michael Miksovic, Christoph Gschwind, Thomas Fusco, Francesco Giovannini, Andrea Giurgiu, Ioana |
| author_facet | Nidd, Michael Miksovic, Christoph Gschwind, Thomas Fusco, Francesco Giovannini, Andrea Giurgiu, Ioana |
| contents | Having access to realistic workloads for a given database instance is extremely important to enable stress and vulnerability testing, as well as to optimize for cost and performance. Recent advances in learned cost models have shown that when enough diverse SQL queries are available, one can effectively and efficiently predict the cost of running a given query against a specific database engine. In this paper, we describe our experience in exploiting modern synthetic data generation techniques, inspired by the generative AI and LLM community, to create high-quality datasets enabling the effective training of such learned cost models. Initial results show that we can improve a learned cost model's predictive accuracy by training it with 45% fewer queries than when using competitive generation approaches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_19807 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Bootstrapping Learned Cost Models with Synthetic SQL Queries Nidd, Michael Miksovic, Christoph Gschwind, Thomas Fusco, Francesco Giovannini, Andrea Giurgiu, Ioana Databases Artificial Intelligence Having access to realistic workloads for a given database instance is extremely important to enable stress and vulnerability testing, as well as to optimize for cost and performance. Recent advances in learned cost models have shown that when enough diverse SQL queries are available, one can effectively and efficiently predict the cost of running a given query against a specific database engine. In this paper, we describe our experience in exploiting modern synthetic data generation techniques, inspired by the generative AI and LLM community, to create high-quality datasets enabling the effective training of such learned cost models. Initial results show that we can improve a learned cost model's predictive accuracy by training it with 45% fewer queries than when using competitive generation approaches. |
| title | Bootstrapping Learned Cost Models with Synthetic SQL Queries |
| topic | Databases Artificial Intelligence |
| url | https://arxiv.org/abs/2508.19807 |