CODEWITHSUNDEEP
variablestype-conversionjulia
Mohammad Saad
Mohammad Saad

Reputation: 2005

How to convert variable type "SentinelArrays" to "Arrays{Float64,n}" in julia

I am trying to optimize my Julia code by making it type-stable. Hence, I tried to declare the variable types in the function headers. But one of the variables has a type of ::SentinelArrays.ChainedVector{Float64,Array{Float64,1}} as shown in the code snippet below.

The code example:

df=CSV.read("text.csv", DataFrame)

a = view(df, :, 1)      
#this has a type of ::SentinelArrays.ChainedVector{Float64,Array{Float64,1}}

b = view(df, :, 2:4)
#while type of this is ::Arrays{Float64,2}

#I would like to pass the type of the arrays in the function.
function calc(a, b::Arrays{Float64,2})
    a+b
end

I tried passing the typeof(a) in the function calc(a::SentinelArrays.ChainedVector{Float64,Array{Float64,1}}, b::Arrays{Float64,2}) however, this throws an error of no method matching. May I know the correct way to assign this type or maybe if can convert this to normal Array{Float64,1}. Please suggest a solution to this issue. Thanks in advance.

Upvotes: 2

Views: 580

Answers (1)

Przemyslaw Szufel
Przemyslaw Szufel

Reputation: 42234

You can just write Array(a) where a is your SentinelArray as here:

julia> u = SentinelArray(rand(1:8,4))
4-element SentinelVector{Int64, Int64, Missing, Vector{Int64}}:
 2
 3
 5
 3

julia> Array(u)
4-element Vector{Union{Missing, Int64}}:
 2
 3
 5
 3

However, normally you would just make the function signature to be something like:

function calc(a, b::AbstractArray{T,2}) where T

because this would work with both those types:

julia> SentinelMatrix{Int64, Int64, Missing, Matrix{Int64}} <: AbstractArray{T,2} where T
true

Upvotes: 2

Related Questions