We are very excited to be working with two fantastic GSoC contributors this year: Yutan Yang (@tarotaroyoung) and Zhenglin Li. Learn more about them at https://t.co/e1Id66yjsk.
How can we automatically and efficiently find performance issues in databases? We propose CERT, a novel black-box technique: https://t.co/eaDaCKP1uw. Our key idea is to test them through the lens of cardinality estimation - a more restrictive query should return fewer rows.
CnosDB becomes the first cloud-native time-series database that supports SQLancer.
By using SQLancer, CnosDB technical team finds and fixes logical errors in product queries, and helps upstream projects build more stable query engines.
https://t.co/YaFqEmZAem
#TimeSeries#CnosDB
Can ChatGPT replace database system testing tools like SQLancer or SQLsmith? An interesting read on an experiment by the @StarRocksLabs community. cc @CelerData https://t.co/S94Dd5qhRn
SQLancer was accepted as a GSoC organization! We have created a "gsoc-2023" Slack channel (https://t.co/dqXFxxSWth). See our wiki for further information and for project ideas (https://t.co/zI15GpxZcW).
We are considering applying for SQLancer to be included as part of Google Summer of Code (GSoC) initiative and are looking for project ideas as well as mentors. Respond if, e.g., you would like to have your database system tested and/or would like to (co-)mentor a contributor!
We already integrated our Query Plan Guidance (QPG) approach into SQLancer (see https://t.co/YG1D3H8Cms). Feel free to give it a try! I also gave a talk on QPG at Tsinghua University, hosted by Yu Jiang, earlier today and received insightful questions and feedback!
Wow, great way to run database test! Although the tweet happened last year, still wanna say congrats to @RiggerManuel to create SQLancer.
Impressive tool to automatically test database in order to find logic bugs during implementation!
#oceanbase#sqlancer#database#test
Our paper "Testing Database Engines via Query Plan Guidance" was accepted at @ICSEconf (3x accepts!). The core idea is to steer automated testing towards unseen query plans for finding logic bugs. We found over 50 unique, previously unknown bugs. Stay tuned for the preprint!
Thanks to @12bytez's contribution, @sqlancer_dbms can now also test @Yugabyte! Great to see that SQLancer already has been useful in locating and preventing bugs (see https://t.co/rKUlNp95kQ).
Thanks to Yisong Han and @Datafuse_Labs for contributing testing support for Databend to @sqlancer_dbms! Hope it will help with preventing bugs before they reach any users. https://t.co/QSzW0s4lo9