Help with error in MixedModels.jl


#1

Hello,

I am trying out MixedModels most basic functioning. I am loading data into Julia using the R package JuliaCall, in order to streamline the needed functionality into my existent R analysis, as described here.

While Douglas Bates’ demo with the kb07 data frame works without a problem, when trying to fit the model
rtlog ~ contrastcstimType + (1 + contrastcstimType | subj) + (1 + contrastc*stimType | stim)

to my data frame (where contrastc and rtlog are numeric, the rest factors),

I get the following error:

MethodError: no method matching getindex(::DataFrames.DataFrame, ::Expr)
Stacktrace:
[1] collect_to!(::Array{AbstractArray{T,1} where T,1}, ::Base.Generator{Array{Any,1},StatsModels.##23#25{DataFrames.DataFrame}}, ::Int64, ::Int64) at ./array.jl:508
[2] collect_to!(::Array{Array{Float64,1},1}, ::Base.Generator{Array{Any,1},StatsModels.##23#25{DataFrames.DataFrame}}, ::Int64, ::Int64) at ./array.jl:518
[3] _collect(::Array{Any,1}, ::Base.Generator{Array{Any,1},StatsModels.##23#25{DataFrames.DataFrame}}, ::Base.EltypeUnknown, ::Base.HasShape) at ./array.jl:489
[4] map(::Function, ::Array{Any,1}) at ./abstractarray.jl:1868
[5] #ModelFrame#22(::Dict{Any,Any}, ::Type{T} where T, ::StatsModels.Terms, ::DataFrames.DataFrame) at /Users/yanivabir/.julia/v0.6/StatsModels/src/modelframe.jl:144
[6] (::Core.#kw#Type)(::Array{Any,1}, ::Type{StatsModels.ModelFrame}, ::StatsModels.Terms, ::DataFrames.DataFrame) at ./:0
[7] #ModelFrame#27(::Array{Any,1},

Any ideas? I am obviously new to Julia, and the lack of any noticeable difference between the kb07 data frame and mine is frustrating…

thanks!


#2

Hi, I’m the maintainer of JuliaCall.
Could you provide the exact code you are running? It is hard to tell where the problem is with just the formula and error messages.


#3

Hi,

I found the issue just now - it’s not to do with JuliaCall, but with MixedModels.jl, which errors when random interaction terms are present in a formula, see this issue: https://github.com/dmbates/MixedModels.jl/issues/100.

Thanks for the response!