eachtarget for dabeled data as stated in documentation

[racinmat]

Hi, I have MLDataPattern.jl v0.5.2, and DataFrames.jl v0.20.2, and I'm trying to reproduce the code in documentation: https://mldatapatternjl.readthedocs.io/en/latest/documentation/targets.html#example-2-symbol-support-for-dataframes-jl
I added

using DataFrames, LearnBase
LearnBase.getobs(df::DataFrame, idx) = df[idx,:]
LearnBase.nobs(df::DataFrame) = nrow(df)

but following code is not working for me, should I add some other method for DataFrames, or is there something else that needs to be done to get this working?

using DataFrames
df = DataFrame(x1 = rand(5), x2 = rand(5), y = [:a,:a,:b,:a,:b])
iter = eachtarget(row->row[1,:y], df)
julia> collect(iter)
ERROR: MethodError: no method matching getindex(::DataFrameRow{DataFrame,DataFrames.Index}, ::Int64, ::Symbol)
Closest candidates are:
  getindex(::DataFrameRow, ::Union{Signed, Symbol, Unsigned}) at C:\Users\racinsky\.julia\packages\DataFrames\S3ZFo\src\dataframerow\dataframerow.jl:94   
  getindex(::DataFrameRow, ::Union{Regex, AbstractArray{T,1} where T, All, Between, InvertedIndex}) at C:\Users\racinsky\.julia\packages\DataFrames\S3ZFo\src\dataframerow\dataframerow.jl:96
  getindex(::DataFrameRow, ::Colon) at C:\Users\racinsky\.julia\packages\DataFrames\S3ZFo\src\dataframerow\dataframerow.jl:98
Stacktrace:
 [1] (::var"#74#75")(::DataFrameRow{DataFrame,DataFrames.Index}) at C:\Projects\SampleLoaders.jl\test\runtests.jl:239
 [2] gettarget at C:\Users\racinsky\.julia\packages\MLDataPattern\nA4s5\src\targets.jl:151 [inlined]
 [3] _gettarget at C:\Users\racinsky\.julia\packages\MLDataPattern\nA4s5\src\targets.jl:144 [inlined]
 [4] (::MLDataPattern.var"#126#127"{var"#74#75"})(::DataSubset{DataFrame,Int64,LearnBase.ObsDim.Undefined}) at .\none:0
 [5] iterate at .\generator.jl:47 [inlined]
 [6] collect(::Base.Generator{ObsView{DataSubset{DataFrame,Int64,LearnBase.ObsDim.Undefined},DataFrame,LearnBase.ObsDim.Undefined},MLDataPattern.var"#126#127"{var"#74#75"}}) at .\array.jl:665
 [7] top-level scope at none:0

Thanks.

[racinmat]

So, I figured out following code does exactly what is expected

using DataFrames
df = DataFrame(x1 = rand(5), x2 = rand(5), y = [:a,:a,:b,:a,:b])
iter = eachtarget(row->row[:y], df)
julia> collect(iter)

but still I guess it would be worth fixing this in docs.

[oxinabox]

I think if you load MLDataUtils.jl it works.

[racinmat]

It seems like not, should I make PR with fix in documentation?

using DataFrames, LearnBase
LearnBase.getobs(df::DataFrame, idx) = df[idx,:]
LearnBase.nobs(df::DataFrame) = nrow(df)
using MLDataPattern, MLDataUtils
df = DataFrame(x1 = rand(5), x2 = rand(5), y = [:a,:a,:b,:a,:b])
iter = eachtarget(row->row[1, :y], df)   # this does not work, collect crashes
collect(iter)

iter = eachtarget(row->row[:y], df)   # this works, collect works
collect(iter)

[oxinabox]

sure