Omega359 commented on PR #9083:
URL:
https://github.com/apache/arrow-datafusion/pull/9083#issuecomment-1919387570
I am uncertain as to whether the unsupported data type errors should always
be internal or exec. I think it may depend on the signature of the function.
For example, from_unixtime is specified as having a signature of
```
BuiltinScalarFunction::FromUnixtime => {
Signature::uniform(1, vec![Int64], self.volatility())
}
```
so if that function got anything other than Int64 it would definitely be
unexpected and thus I would think that internal_err might be appropriate there.
However, to_timestamp(_xxx) methods are defined as
```
BuiltinScalarFunction::ToTimestamp
| BuiltinScalarFunction::ToTimestampSeconds
| BuiltinScalarFunction::ToTimestampMillis
| BuiltinScalarFunction::ToTimestampMicros
| BuiltinScalarFunction::ToTimestampNanos => {
Signature::variadic_any(self.volatility())
}
```
So it would be expected that the impl would have to handle and return
unacceptable types (thus exec_err)
--
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]
