Is your feature request related to a problem or challenge?

Math functions in datafusion-functions do not support Decimal128, Decimal256 directly, always casting the input to floats.

There is merged support for log function in #17140. This case is simpler since the function has exactly two arguments (log has one or two).

Describe the solution you'd like

Implement type cases for decimals using a generic calculate_binary_math macro.

Describe alternatives you've considered

No response

Additional context

No response