cmudig/solas

data type inference should propagate to new columns

willeppy opened this issue · 2 comments

df = data.cars() # Cylinders is inferred as nominal
df.new_Cylinders = df.Cylinders * 8 # changes type of Cylinders to quant but not new_Cylinders

In above example, when we update the types of Cylinders it should also apply to new_Cylinders since it is derived. Need to figure out when Lux sets the type for new columns, this might be easy fix

I added some codes to the __setitem__ function of the LuxDataFrame in which the data type of the new value is copied

fixed in pr