Looks the interpretation of not specifying an intercept changes depending on whether you is ModelFrame or apply_schema, see the following code snippet
using StatsModels, DataFrames
f = @formula(y ~ x)
df = DataFrame(y = [1, 2, 3], x = [4, 5, 6])
mf = ModelFrame(f, df)
mf.f #has an intercept
apply_schema(f, schema(f, df)) #no intercept
Also since ModelFrame is not encouraged, it's a bit strange to have the categorical variable contrasts locked behind that function, unless there's a simple way to use them with apply_schema?
Looks the interpretation of not specifying an intercept changes depending on whether you is
ModelFrameorapply_schema, see the following code snippetAlso since
ModelFrameis not encouraged, it's a bit strange to have the categorical variable contrasts locked behind that function, unless there's a simple way to use them withapply_schema?