CODEWITHSUNDEEP
listjulia
Mohammad Saad
Mohammad Saad

Reputation: 2005

How to covert an array of type Float to Any in Julia?

I would like to convert an array to a python like list in Julia.

To create an array of type Any one may initialize an empty array with this syntax array = []. However if I want to perform conversion, i.e. lets take an array of type Array(Float64,1) to type Any, what would be the correct approach?

Or If there are any alternate approaches to create a list in Julia?

My approach to create a function which takes an array and convert it to type any:

function list(x)
    x = convert(Any, x)
    return x
end

x_test = [0.19, 0.03, 0.27]
t1 = typeof(x_test)
println("type of x_test is: $t1")
x_test = list(x_test)
t2 = typeof(x_test)
println("type of x_test is: $t2")

Output:
type of x_test is: Array{Float64,1}
type of x_test is: Array{Float64,1}

Please suggest a method or solution to achieve this conversion task.

Thanks.

Upvotes: 3

Views: 918

Answers (3)

Benoit Pasquier
Benoit Pasquier

Reputation: 2995

You were not far, you just need to convert to an array of Any:

julia> convert(Array{Any}, x_test)
3-element Vector{Any}:
 0.19
 0.03
 0.27

But as others have said, it's not a good idea to hide type information in general because it will just slow things down.

Upvotes: 1

Przemyslaw Szufel
Przemyslaw Szufel

Reputation: 42214

The shortest form is Vector{Any}(a) as in code here:

julia> a=[1,2,3]
3-element Vector{Int64}:
 1
 2
 3

julia> Vector{Any}(a)
3-element Vector{Any}:
 1
 2
 3

However, if you want to be able to hold in a copy of a other elements such as Strings you will be much more efficient by strictly telling that:

julia> b = Vector{Union{eltype(a),String}}(a)
3-element Vector{Union{Int64, String}}:
 1
 2
 3

julia> push!(b,"jan")
4-element Vector{Union{Int64, String}}:
 1
 2
 3
  "jan"

Upvotes: 5

Nils Gudat
Nils Gudat

Reputation: 13800

You can do:

julia> list(x) = Any[i for i ∈ x]
list (generic function with 1 method)

julia> list([0.19, 0.03, 0.25])
3-element Vector{Any}:
 0.19
 0.03
 0.25

But as Oscar says in his comments, why would you ever want to do that? It's true that Python performance often suffers because of a lack of type information, that doesn't mean Julia becomes "like Python" if you deliberately prevent the compiler from optimizing (it will become a lot slower though in almost all cases!)

Upvotes: 6

Related Questions