Add kernel to quickly compute comparisons on Arrays

[alamb]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
DataFusion (and likely other libraries) often wants to implement its own fast comparison functions

For example, comparing the contents of an array element by element to form an output BooleanArray
https://github.com/apache/arrow-datafusion/blob/7617d78809d4ff5bde31142e0744c70024e40635/datafusion/physical-expr/src/expressions/in_list.rs#L50

the arity::unary is close to what I want: https://docs.rs/arrow/17.0.0/arrow/compute/kernels/arity/index.html but it can not produce a BooleanArray (because BooleanArray is not a PrimitiveArray)

fn in_list_primitive<T: ArrowPrimitiveType>(
    array: &PrimitiveArray<T>,
    values: &[<T as ArrowPrimitiveType>::Native],
) -> Result<BooleanArray> {
    Ok(arity::unary(array, |x| values.contains(&x)))
}

Describe the solution you'd like
I would like a function such as unary_cmp that I could call like:

fn in_list_primitive<T: ArrowPrimitiveType>(
    array: &PrimitiveArray<T>,
    values: &[<T as ArrowPrimitiveType>::Native],
) -> Result<BooleanArray> {
    Ok(arity::unary_cmp(array, |x| values.contains(&x)))
}

There is a proposed implementation: apache/datafusion#2826

Describe alternatives you've considered

Additional context
Add any other context or screenshots about the feature request here.

[viirya]

Instead of copying unary, seems we can reuse exist compare code like #1991