maropu commented on issue #27872: [SPARK-31115][SQL] Detect known Janino bug janino-compiler/janino#113 and apply workaround automatically as a fail-back via avoid using switch statement in generated code URL: https://github.com/apache/spark/pull/27872#issuecomment-599044117 I think the option 4 looks fine to me. btw, splitting large code into pieces in `switch` is a solution for this issue? Additionally, we need to replace `switch` with `if`? > Modify ExpandExec to check the number of operations in for statement, and use if ~ else if when the number of operations exceed the threshold. This should be ideally checking the length of offset but it would be weird if Spark does it, so count the lines blindly. Performance regression may happen in some cases where it can run with switch but due to blind count it runs with if ~ else if, but the case wouldn't be common. I just want to know the actual performance numbers of this approach. I think splitting large code into small parts might improve performance. > I have one, but I cannot share since the query is from actual customer. If you're OK with just generated code, I've attached the file in Janino issue janino-compiler/janino#113. To reproduce the issue, could you build the simple query that you can show us based on your private customer's query? I think the query can make us understood more for the issue.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
