error in assigning NA to Float64 in a DataFrame

[Min-Woong Sohn]

I am trying to work out the following example:

using DataFrames, GLM

N = 1000

data = DataFrame(X=randn(N),Y=randn(N))

beta_0 = 0

beta_1 = 1

index = (beta_0 + beta_1*data[:X])

probability = cdf(Normal(0,1),index)

D = zeros(N)

for i=1:N

    D[i]=rand(Bernoulli(probability[i]))

end

data[:D]=D

data2 = data

data2[data2[:D] .== 0.0, :Y] = NA

data2

data3 = DataFrame(rand(MvNormal([0,0.],[1 .5;.5 1]),N)')

names!(data3,[:X,:Y])

mean(convert(Array,data3),1)

cov(convert(Array,data3))

index = (beta_0 + beta_1*data3[:X])

probability = cdf(Normal(0,1),index)

D = zeros(N)

for i=1:N

    D[i]=rand(Bernoulli(probability[i]))

end

data3[:D]=D

data4 = data3

data4[data4[:D] .== 0.0, :Y] = NA

Lines 14 and 29 are exactly the same but the first one works fine and the second one gives an error: LoadError: MethodError: `convert` has no method matching convert(::Type{Float64}, ::DataArrays.NAtype)

Does anyone know what's happening here?

[Alex Williams]

The problem is that data4[:Y] is an Array, not a DataArray... so you can't add NA values to it. I think it works if you add the following command right before the last line:

data4[:Y] = convert(DataArray,data4[:Y])

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-stats...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Min-Woong Sohn]

Thanks a bunch. It works now.