pepijnve commented on code in PR #18191:
URL: https://github.com/apache/datafusion/pull/18191#discussion_r2448651428
##########
datafusion/functions/src/core/nvl2.rs:
##########
@@ -95,8 +96,71 @@ impl ScalarUDFImpl for NVL2Func {
Ok(arg_types[1].clone())
}
+ fn return_field_from_args(&self, args: ReturnFieldArgs) ->
Result<FieldRef> {
+ let nullable =
+ args.arg_fields[1].is_nullable() ||
args.arg_fields[2].is_nullable();
+ let return_type = args.arg_fields[1].data_type().clone();
+ Ok(Field::new(self.name(), return_type, nullable).into())
+ }
+
fn invoke_with_args(&self, args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
- nvl2_func(&args.args)
+ let [test, if_non_null, if_null] = take_function_args(self.name(),
args.args)?;
Review Comment:
In #17357 the author chose to make `invoke_with_args` return an internal
error instead of retaining the implementation. Would we want to do the same
here?
I'm a bit on the fence myself. On the one hand, this is effectively dead
code for most users. On the other hand, raising an error here may cause
breakage for users who have customised their optimiser passes and are not doing
simplification. No idea if anyone actually does that.
--
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]
