In some datasets, we may find that a float column has a mixture of precisions. E.g. half the column is rounded to the nearest 0.5, the rest to the nearest 0.00001. It would be good to be able to produce this, I can't think of a good way to do it off the top of my head, other than creating an augmented indicator for each detected / specified rounding scheme and applying accordingly.
This is a nice to have feature, and I would suggest that data of this type is poorly managed and a decision should be made to standardise the precision of the feature.
In some datasets, we may find that a float column has a mixture of precisions. E.g. half the column is rounded to the nearest 0.5, the rest to the nearest 0.00001. It would be good to be able to produce this, I can't think of a good way to do it off the top of my head, other than creating an augmented indicator for each detected / specified rounding scheme and applying accordingly.
This is a nice to have feature, and I would suggest that data of this type is poorly managed and a decision should be made to standardise the precision of the feature.