lift missings in response, fitted, residuals?

[kolia]

Currently, data rows that containmissings in the model inputs are silently dropped, so the output vector is shorter than the input by however many rows had missings. I think it would be more user-friendly to lift missings and return a vector with the same number of elements as the input has rows, with missings at indices that had missings in the inputs.

[kleinschmidt]

THis is a long-standing issue in statsmodels land: JuliaStats/StatsModels.jl#202 (and see this WIP PR from @palday JuliaStats/StatsModels.jl#153) and should probably be handled at that level. But it's definitely on the road map there.

[ararslan]

Not the same but see also JuliaStats/GLM.jl#496