quantileExact Function Returns nan for Empty Float32/Float64 Sets

Why it matters

Previously, the behavior of the quantileExact function on empty floating-point datasets was undefined or inconsistent. This update provides a clear and consistent result of nan for such cases, improving reliability and clarity for users performing exact quantile calculations on empty sets.

How to use it

Use the quantileExact function as usual in your queries. When the input set (for Float32 or Float64 data types) is empty, the function will return nan automatically without any additional configuration.