factor levels for ddf variables
Opened this issue · 0 comments
hafen commented
It is important for categorical variables in a distributed data frame to be able to be treated as factors. We have the problem of not knowing all the possible factor levels without looking at the entire data. But when we call updateAttributes, we get frequency tables for all categorical variables. We can extract the factor levels from this - we just need to make this part of how ddfs are handled.