Decimal multiplication / addition operations are lossy

[Dandandan]

Describe the bug
e.g multiply_decimal

pub(crate) fn multiply_decimal(
    left: &Decimal128Array,
    right: &Decimal128Array,
) -> Result<Decimal128Array> {
    let divide = 10_i128.pow(left.scale() as u32);
    let array = multiply(left, right)?;
    let array = divide_scalar(&array, divide)?
        .with_precision_and_scale(left.precision(), left.scale())?;
    Ok(array)
}

Loses information, as integer division is used instead of increasing the precision / scale.
The same applies for addition as well.

To Reproduce
Steps to reproduce the behavior:

Expected behavior
A clear and concise description of what you expected to happen.

Additional context
Add any other context about the problem here.

[tustvold]

I wonder if part of the problem is that currently mathematical operations are not expected to alter the type of data, whereas this would?

[Dandandan]

That might be part of the problem / why this issue bubbles up (for this implementation at least).
I don't think it would be strictly necessary to keep this property, as the datatype is dynamic.

We should probably have to change some type calculation (get_type etc.) in DataFusion to correctly give the outcome types though given the decimal operations.

[alamb]

cc @liukun4515 who may have opinions on this matter