Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.21674 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909985879883776 |
|---|---|
| author | Ananthakrishnan, Haritha Kokel, Harsha Sikes, Kelsey Bhattacharjya, Debarun Katz, Michael Sohrabi, Shirin Srinivas, Kavitha |
| author_facet | Ananthakrishnan, Haritha Kokel, Harsha Sikes, Kelsey Bhattacharjya, Debarun Katz, Michael Sohrabi, Shirin Srinivas, Kavitha |
| contents | We introduce QueryGym, an interactive environment for building, testing, and evaluating LLM-based query planning agents. Existing frameworks often tie agents to specific query language dialects or obscure their reasoning; QueryGym instead requires agents to construct explicit sequences of relational algebra operations, ensuring engine-agnostic evaluation and transparent step-by-step planning. The environment is implemented as a Gymnasium interface that supplies observations -- including schema details, intermediate results, and execution feedback -- and receives actions that represent database exploration (e.g., previewing tables, sampling column values, retrieving unique values) as well as relational algebra operations (e.g., filter, project, join). We detail the motivation and the design of the environment. In the demo, we showcase the utility of the environment by contrasting it with contemporary LLMs that query databases. QueryGym serves as a practical testbed for research in error remediation, transparency, and reinforcement learning for query generation. For the associated demo, see https://ibm.biz/QueryGym. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_21674 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | QueryGym: Step-by-Step Interaction with Relational Databases Ananthakrishnan, Haritha Kokel, Harsha Sikes, Kelsey Bhattacharjya, Debarun Katz, Michael Sohrabi, Shirin Srinivas, Kavitha Databases Artificial Intelligence We introduce QueryGym, an interactive environment for building, testing, and evaluating LLM-based query planning agents. Existing frameworks often tie agents to specific query language dialects or obscure their reasoning; QueryGym instead requires agents to construct explicit sequences of relational algebra operations, ensuring engine-agnostic evaluation and transparent step-by-step planning. The environment is implemented as a Gymnasium interface that supplies observations -- including schema details, intermediate results, and execution feedback -- and receives actions that represent database exploration (e.g., previewing tables, sampling column values, retrieving unique values) as well as relational algebra operations (e.g., filter, project, join). We detail the motivation and the design of the environment. In the demo, we showcase the utility of the environment by contrasting it with contemporary LLMs that query databases. QueryGym serves as a practical testbed for research in error remediation, transparency, and reinforcement learning for query generation. For the associated demo, see https://ibm.biz/QueryGym. |
| title | QueryGym: Step-by-Step Interaction with Relational Databases |
| topic | Databases Artificial Intelligence |
| url | https://arxiv.org/abs/2509.21674 |