I am using nuniuque method in both pandas and vaex to count and/or compare float values in aggregated groups.
In pandas though, when filtering out rows I use np.isclose method to be able to set relative tolerance and avoid errors in comparison but I do not know how to do the same in the grouped dataframe (in pandas and vaex).Reading this questions pointed me that rounding numbers to the same precision is not a solution here.
How can I safely compare numbers with up to 4 digits after floating point in grouped pandas or vaex dataframe considering that I am not calculating those numbers but downloading them from an SQL table.
