Akshat-Jain commented on PR #16729: URL: https://github.com/apache/druid/pull/16729#issuecomment-2225538746
> Alongside, I think its better to not use the .q and .e files to run the normal test cases. It checks only the output, which is expected for drill tests. In general we would loose the capability of the asserting other aspects of planning and execution (for example - test cases under shuffle columns) @sreemanamala It's intentional. I wanted to add a bunch of tests where we validate that the outputs of MSQ and sql-native engine are the same. We have a very nice wiring already for it in the Drill tests, so I re-used that. We have a `MSQWindowTest` where more specific tests with more detailed validations should go - but that's MSQ specific. For now, I was hoping to validate that queries lead to same result with both engines. I expect to add a LOT of tests in the DruidWindowQueryTest file that I added, that would run for both engines. I feel we should have more tests that run for both engines, even if they verify only the outputs for now. But to your point, I will add more specific tests to the MSQ specific test file as well for more detailed assertions. I'm kinda hoping to revamp the rest of the logic in the next 2-3 PRs, detailed tests would make more sense then. Hope that sounds good! :) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
