If I have a Float64 vector Y and an integer vector x, for instance x=rand(1:1000, 500), is there an elegant way to pull the elements of Y at non-x entries? So far I have tried Y[findall([i ∉ x for i in 1:1000])]. This works, but coming from R, I was hoping to do something like Y[.!x] or Y[!x], which both throw errors. I would like to refrain from a package like DataFrames, but if this is not possible I understand.
Thanks in advance.
Use Not from InvertedIndices (this also gets imported with DataFrames).
In your case this is Y[Not(x)], see the code below:
julia> using InvertedIndices # or using DataFrames
julia> Y = collect(1:0.5:4)
7-element Vector{Float64}:
1.0
1.5
2.0
2.5
3.0
3.5
4.0
julia> x=rand(1:7, 3)
3-element Vector{Int64}:
3
2
6
julia> Y[Not(x)]
4-element Vector{Float64}:
1.0
2.5
3.0
4.0
Since the question explicitly asked for a solution that does not rely on packages outside of the standard library here is an alternative to Przemyslaw Szufel's solution:
Y[∉(x).(1:length(Y))]
Here we use the partially applied form of ∉. From the documentation for in:
in(collection)
∈(collection)Create a function that checks whether its argument is
in collection, i.e. a function equivalent toy -> y in collection.
The same thing can be written in a few different ways, e.g. Y[eachindex(Y) .∉ Ref(x)] (works for this case but you should understand eachindex and have a look at LinearIndices and CartesianIndices).
An important thing to note is that these solutions do not perform well when x is large. To improve performance a Set can be created from x. Example:
Y[∉(Set(x)).(eachindex(Y))]