Julia Lang: Error? Initializing Nullable Arrays - arrays

Question

I am trying to initialize an array of Nullable Integers with a given size. My intend is to add elements to the array and to be able to count the number of Nullable elements in an array.
However I noticed the following behavior. Running the following code:
function main()
arr = Array{Nullable{Int}}(10)
res = 0
for i in 1:10
if isnull(arr[i])
res += 1
end
end
res
end
is expected to return 10, however the program returns different values almost every time that I run the code. Initializing just the array:
function main()
arr = Array{Nullable{Int}}(10)
end
this is the result:
Nullable{Int64}(7142821636481377634)
Nullable{Int64}(8803814271447229814)
Nullable{Int64}(7935455170894001012)
Nullable{Int64}(8247625210579135584)
Nullable{Int64}(7815275285807655200)
Nullable{Int64}()
Nullable{Int64}()
Nullable{Int64}()
Nullable{Int64}()
Nullable{Int64}()
Here one can see that not all elements are in fact not Null. Which would explain why the function isnull() does not recognize them as Null objects.
This is the desired effect:
function main()
arr = Array{Nullable{Int}}(10)
for i in 1:10
arr[i] = Nullable{Int}()
end
println(arr)
end
which indeed returns an array of Nullable elements. However, it seems a bit tedious to run this code for initialization.
Any ideas how this can be avoided?

Answer 1

In Julia 0.6 you can use arr = fill(Nullable{Int64}(),10).
Note however that Nullables are depreciated in Julia 0.7 (https://github.com/JuliaLang/julia/blob/master/NEWS.md), consider to use the Nothing or Missing types instead:
julia> vector_with_nothing_or_int = Vector{Union{Int64, Nothing}}(nothing,10)
10-element Array{Union{Nothing, Int64},1}:
nothing
nothing
[...]
julia> vector_with_missing_or_int = Vector{Union{Int64, Missing}}(missing,10)
10-element Array{Union{Missing, Int64},1}:
missing
missing
[...]
Check if a element is nothing or missing can be done with:
julia> vector_with_nothing_or_int[1] == nothing
true
julia> ismissing(vector_with_missing_or_int[1])
true
The Missing type also works in Julia 0.6 if you load the module Missings. This is the corresponding code for Julia 0.6:
using Missings
vector_with_missing_or_int = Vector{Union{Int64, Missing}}(10)
vector_with_missing_or_int .= missing
vector_with_nothing_or_int = Vector{Union{Int64, Nothing}}(10)
vector_with_nothing_or_int .= nothing
The operator .= replaces every element of the array with the element of the right. All these examples also work for multi-dimensional arrays if you replace Vector by Array when you specify also the number of dimensions.