NaN if inputs have only 1 element

[wjholden]

https://github.com/stefftaelman/Xicor.jl/blob/f6c9155ef2efcda7e2ef8ea3a1e9b6f434a2cad8/src/Xicor.jl#L16

I propose a small check here that the input vectors have at least two elements each. The program returns a NaN statistic if n=1 (with our without ties). If n=0, then we get a meaningless statistic of 1.0 with no ties and a confusing error with ties.

I recommend an explicit ArgumentError or DomainError to explicitly tell the user that they cannot use this function on inputs with n<2.

[wjholden]

julia> xicor([1], [1]; noties=true)
NaN

julia> xicor([1], [1]; noties=false)
NaN

julia> xicor([], []; noties=true)
1.0

julia> xicor([], []; noties=false)
ERROR: ArgumentError: reducing over an empty collection is not allowed; consider supplying `init` to the reducer
Stacktrace:
  [1] _empty_reduce_error()
    @ Base .\reduce.jl:319
  [2] mapreduce_empty(f::Function, op::Function, T::Type)
    @ Base .\reduce.jl:321
  [3] reduce_empty(op::Base.MappingRF{Xicor.var"#4#7"{Int64}, typeof(Base.add_sum)}, ::Type{Int64})
    @ Base .\reduce.jl:358
  [4] reduce_empty_iter
    @ .\reduce.jl:381 [inlined]
  [5] mapreduce_empty_iter(f::Function, op::Function, itr::Vector{Int64}, ItrEltype::Base.HasEltype)
    @ Base .\reduce.jl:377
  [6] _mapreduce
    @ .\reduce.jl:429 [inlined]
  [7] _mapreduce_dim
    @ .\reducedim.jl:337 [inlined]
  [8] mapreduce
    @ .\reducedim.jl:329 [inlined]
  [9] _sum
    @ .\reducedim.jl:987 [inlined]
 [10] sum
    @ .\reducedim.jl:983 [inlined]
 [11] xicor(X::Vector{Any}, Y::Vector{Any}; rank::typeof(StatsBase.denserank), noties::Bool)
    @ Xicor C:\Users\wjhol\Documents\GitHub\Xicor.jl\src\Xicor.jl:27
 [12] top-level scope
    @ REPL[5]:1