BEAVER: Enterprise benchmark for LLM Text-to-SQL from private data warehouses

teleforce2 pts0 comments

-->

-->

-->

-->

-->

-->

BEAVER: An Enterprise Benchmark for Text-to-SQL

🦫 BEAVER: An Enterprise Benchmark for Text-to-SQL

Peter Baile Chen1,

Devin Yang1,

Weiyue Li2,

Fabian Wenz1,3,

Yi Zhang4,

Nesime Tatbul1,5,

Michael Cafarella1,

Çağatay Demiralp1,6,

Michael Stonebraker1

1MIT, 2Harvard University,<br>3TU Munich,<br>4Greenshoe, Inc.,<br>5Intel,<br>6AWS AI Labs

arXiv

Dataset

Code

Overview

Leaderboard

Leaderboard

Submission Instructions

How to Submit:

Please send an email to peterbc@mit.edu, along with your method name, a brief description of the method, and, optionally, a link to your paper or codebase. We will follow up with detailed instructions.

Method

Model

Search

Rank<br>Submission Date<br>Method<br>Model<br>Execution Accuracy

Subtask Metrics Across Settings

Citation

If you find our data, code, or the paper helpful, please cite the<br>paper:

@article{chen2024beaver,<br>title={BEAVER: an enterprise benchmark for text-to-sql},<br>author={Chen, Peter Baile and Yang, Devin and Li, Weiyue and Wenz, Fabian and Zhang, Yi and Tatbul, Nesime and Cafarella, Michael and Demiralp, {\c{C}}a{\u{g}}atay and Stonebraker, Michael},<br>journal={arXiv preprint arXiv:2409.02038},<br>year={2024}

Copy

BEAVER is a large-scale enterprise text-to-SQL dataset containing 9128 queries spanning 812 tables across 19 diverse domains. Of these, 7978 queries are publicly released, while the remaining portion is held out as a private test set. Queries and databases were collected from private organizations.

To facilitate fine-grained evaluation and analysis, we provide

annotations for five subtasks: multi-table retrieval, join key detection, column mapping, domain knowledge extraction, and query decomposition

three categories of queries: complex queries without domain knowledge, domain-specific queries with minimal complexity, and domain-specific complex queries

Example data

Representative BEAVER tasks with question, SQL, and subtask annotations.

Task id

db<br>source

Question

SQL

Multi-Table retrieval

Tables used in SQL

Join Keys

Connections among used tables

Column Mapping

Mapping from question phrases to table columns(s)

Domain knowledge

Domain-specific predictates used in SQL

Subquery Decomposition

Decomposition of SQL into simpler sub-queries

Usage

We have created a unified MySQL version for our dataset. A free<br>MySQL installation can be found<br>here. After<br>the installation, import the MySQL dump files from the google<br>drive to your local MySQL databases using

mysql -u root -p

To execute a SQL statement, you can either log in to the MySQL<br>interface or you can do it via<br>mysql-connector-python.

If you want to use the Oracle version of<br>DW queries, you can download the free oracle<br>database and import the CSVs.

-->

Changelog

-->

Citation

If you find our data, code, or the paper helpful, please cite the<br>paper:

article{chen2024beaver,<br>title={BEAVER: an enterprise benchmark for text-to-sql},<br>author={Chen, Peter Baile and Yang, Devin and Li, Weiyue and Wenz, Fabian and Zhang, Yi and Tatbul, Nesime and Cafarella, Michael and Demiralp, {\c{C}}a{\u{g}}atay and Stonebraker, Michael},<br>journal={arXiv preprint arXiv:2409.02038},<br>year={2024}

Copy

queries beaver mysql enterprise text michael

Related Articles