jorgecarleitao commented on pull request #8032: URL: https://github.com/apache/arrow/pull/8032#issuecomment-686533676
I agree with you, @andygrove . * Built-in functions are not part of the registry as they are exposed directly from the `prelude`. So, maybe `udf_registry` could be more explicit about this. * I thought about `FunctionRegistry` being used for both udfs and UDAFs (user-defined aggregate function, à la [spark](https://spark.apache.org/docs/latest/sql-ref-functions-udf-aggregate.html)), so that a user does not have to call `df.registry()` or `df.aggregate_registry()`, they just need to remember about one. * I thought about `FunctionRegistry::udf` and `FunctionRegistry::udaf`, just because short names make it easier to read statements. But, as always, it is a trade-off between short vs understanding. I have no strong opinions about naming nor UX here: I will implement whatever you agree upon :-). My main concern was to fix the data type thing :P ---------------------------------------------------------------- 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: us...@infra.apache.org
