A Demo of SQL Rewrite.

---

Overview • Installation • Datasets • Issues • Citation •

What's New?

• May 2022: First Commit

Overview

evolveRewrite is a learned SQL transformation tool, which takes as input a SQL query and corresponding statistics (e.g., schema, #-table rows), finds the optimal rewrite sequence and outputs an optimized rewritten query. Currently evolveRewrite is developled based on Calcite.

Installation

Note: 仅需要安装JAVA环境就可运行项目，早期的JDK版本在Calcite上有编码问题，所以建议使用JDK11以上版本。 点击查看安装文档

What Can You Do via evolveRewrite?

Code Usage

• 可在Interllij IDEA等IDE中直接运行test。
• 当前测试数据基于TPCH,如需修改为其它测试数据可修改src/main/schema.json中json格式的schema
• 修改src/test文件中的testSql进行语句测试

Use Api

项目中带了HttpServer,可直接运行nohup /root/jdk-15/bin/java -jar {{输入项目路径}}/rewriter_java.jar --server.port=6336 &，端口为6336

• api: /rewriter POST {sql: "select....", schema: {....}}

Datasets

For the online version, we support datasets with different legal schema. By default we adopt TPC-H schema in evolveRewrite.

Issues

Major improvement/enhancement in future.

• add custom rules
• update the RA2SQL script (e.g., avoiding useless changes from the origin SQL)
• support cost models of mainstream databases
• verify rewriting equivalence
• add more rewriting information in the front-end pages
• test under more real scenarios (welcome to contact us:))

Citation

Please cite our paper if you use SQL-Rewriter in your work

@article{DBLP:journals/pvldb/ZhouLCF21,
  author    = {Xuanhe Zhou and
               Guoliang Li and
               Chengliang Chai and
               Jianhua Feng},
  title     = {A Learned Query Rewrite System using Monte Carlo Tree Search},
  journal   = {Proc. {VLDB} Endow.},
  volume    = {15},
  number    = {1},
  pages     = {46--58},
  year      = {2021},
}

Contributors

We thank all the contributors to this project, more contributors are welcome!