What is the most convinient format for the population sampels?

[scheidan]

Currently we return the samples as Vector{Vector} . This makes it not very nice to work with the samples and most user would probably run

stack(result.samples)

to get a matrix n_parameter x n_samples matrix back. Should we return the result directly in this format?

Other ideas:

• DimensionalData.jl looks but may be an overkill great:

https://rafaqz.github.io/DimensionalData.jl/stable/basics

using DimensionalData
A = DimArray(stack(res_1.population),
             (para = ["p_$i" for i in 1:2], sample = ["s_$i" for i in 1:5000]),
             name = "population sample")

A[1:2, 4:6]
A[para=2, sample=:]
Array(A)

• AxisKeys.jl seems lighter
  https://github.com/mcabbott/AxisKeys.jl

using AxisKeys

B = KeyedArray(stack(res_1.population),;
               para = [Symbol("p_$i") for i in 1:2],
               iter = 1:5000)

B[1, 3:10]
B[iter = 2:40]

B(:p_1)
B(:p_1, :)
B(:p_1, 10:20)