Open nmprista opened 1 year ago
I think that code run on an data object is a track of changes unless you need to treat rows that have changes differently in estimation. This case seems a slippery slope to me as the function would behave differently on the same data that is reached in different ways.
We need a way to track changes in the data. One possibility is to add a new colum that track changes on the
RDBESdataobject.
Another is to add a columns that track changes introduced by generateMissingSSRows, generateZerosUsingSL, generateNAsUsingSL.