jayzhan211 opened a new issue, #9518: URL: https://github.com/apache/arrow-datafusion/issues/9518
### Is your feature request related to a problem or challenge? In #9504, we have an issue that `select array_ndims([[[[[[[[[[[[[[[[[[[[[1]]]]]]]]]]]]]]]]]]]]])` is taking so long time compare with the main branch. I found out that we always calculate the arguments type every time, which is really costly for the expression like above. https://github.com/apache/arrow-datafusion/blob/5537572820977b38719e2253f601a159deef5bc6/datafusion/expr/src/expr_schema.rs#L141-L156 UDF now always go through `return_type_from_exprs` https://github.com/apache/arrow-datafusion/blob/5537572820977b38719e2253f601a159deef5bc6/datafusion/expr/src/udf.rs#L306-L316 Always recompute types here. I test it locally that every avoid the re-computation, at least this query gets super fast as the main branch does. ### Describe the solution you'd like Ideally, we should cache the result for the same arguments and schema pair, but that need to check the matching first. I think we can just add another argument `args_types: &[DataType]` for `return_type_from_exprs`. If the `args_types` is empty, we calculate the types, otherwise we skip it. ### Describe alternatives you've considered Other better design ### Additional context _No response_ -- 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]
