Describe the bug, including details regarding any error messages, version, and platform.

Below codes calculate the decimal type's precision and scale.

It seems different from our Redshift’s decimal promotion rules, is there any relevant background here?

scale = max(4, s1 + p2 - s2 + 1)
precision = p1 - s1 + s2 + scale

It may cause the decimal divide results different from normal expression calculate.

Component(s)

C++ - Gandiva