DataType::is_numeric should match the is_numeric function in Datafusion.

[HaoYang670]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
In datafusion, we have:

/// Determine if a DataType is signed numeric or not
pub fn is_signed_numeric(dt: &DataType) -> bool {
    matches!(
        dt,
        DataType::Int8
            | DataType::Int16
            | DataType::Int32
            | DataType::Int64
            | DataType::Float16
            | DataType::Float32
            | DataType::Float64
            | DataType::Decimal128(_, _)
    )
}

/// Determine if a DataType is numeric or not
pub fn is_numeric(dt: &DataType) -> bool {
    is_signed_numeric(dt)
        || matches!(
            dt,
            DataType::UInt8 | DataType::UInt16 | DataType::UInt32 | DataType::UInt64
        )
}

We should update DataType::is_numeric to contain these constructors so that we could remove the redundant is_numeric in datafusion (apache/datafusion#1613)

Describe the solution you'd like

Describe alternatives you've considered

Additional context

[HaoYang670]

Hi @alamb, any background of why we don't support Float16 in the cast kernel and the ipc writer?

[liukun4515]

Because lack of some kernel function for some data type which may be not used by the upstream project or other user, the checker of data type maybe duplicated and redundant.

[HaoYang670]

Thank you @liukun4515.
I guess we cannot support Float16 as a numeric type so far because lexical_core::FromLexical doesn't implement for half::f16.

Then, the next question is that should we update the is_numeric function in datafusion to match the arrow-rs. (aka removing float16 and decimal128)? Or should we rename it if it provides different functionality from DataType::is_numeric?

[alamb]

I think f16 support was relatively recent -- added by @Jimexist in #888 . I don't think there was any specific reason that f16 doesn't have as wide support as the other types.

I would be in favor of having arrow-rs and datafusion match. I think decimal128 is a critical usecase for datafusion - as at least @liukun4515 and his team are using it heavily.

I don't know of anyone using f16 at the moment (but maybe there are) so I don't think it is as important