CircularArrayBuffers.jl is a small package to wrap an AbstractArray as a buffer along the last dimension. The main benefit compared to CircularBuffer in DataStructures.jl is that the view of consecutive elements is a SubArray.
julia> using CircularArrayBuffers
julia> names(CircularArrayBuffers)
5-element Array{Symbol,1}:
:CircularArrayBuffer
:CircularArrayBuffers
:CircularVectorBuffer
:capacity
:isfull
julia> a = CircularArrayBuffer(rand(2,3))
2×3 CircularArrayBuffer{Float64,2}:
0.0510714 0.0260738 0.0245707
0.856257 0.571643 0.0189365
julia> b = CircularArrayBuffer{Float64}(2,3)
2×0 CircularArrayBuffer{Float64,2}
julia> push!(b, rand(2))
2×1 CircularArrayBuffer{Float64,2}:
0.4215856115651755
0.5485806794787502
julia> push!(b, rand(2))
2×2 CircularArrayBuffer{Float64,2}:
0.421586 0.640501
0.548581 0.774729
julia> push!(b, rand(2))
2×3 CircularArrayBuffer{Float64,2}:
0.421586 0.640501 0.653054
0.548581 0.774729 0.902611
julia> push!(b, rand(2))
2×3 CircularArrayBuffer{Float64,2}:
0.640501 0.653054 0.640373
0.774729 0.902611 0.227435
julia> pop!(b)
2-element view(::CircularArrayBuffer{Float64,2}, :, 3) with eltype Float64:
0.6403725468830439
0.22743495787074597
julia> b
2×2 CircularArrayBuffer{Float64,2}:
0.640501 0.653054
0.774729 0.902611
julia> size(b)
(2, 2)
julia> capacity(b)
3
julia> isfull(b)
false
julia> push!(b, rand(2))
2×3 CircularArrayBuffer{Float64,2}:
0.640501 0.653054 0.885887
0.774729 0.902611 0.0332439
julia> isfull(b)
true
julia> eltype(b)
Float64