How to get the index of the rows with non-NaN values for multiple vectors?

General Usage
, ,
1

I have 4 vectors of data: A, B, C, D. My goal is to get the index of the rows where all 4 values being non-NaN.

Here is my code:
Ind1 = intersect.(.!isnan.(A), .!isnan.(B), .!isnan.(C), .!isnan.(D));

Why doesn’t it work? The results are like the below:
Vector{Bool}[[0], [0], [0], [0], [0], [0], [0], [0], [0], [0]];

What is the correct way to do this? Many thanks!

2

Your intersection does actually have the information you want, but it’s slightly weird as you are intersecting several numbers, Bool, and numbers are iterable with one element.

There are many other ways, perhaps some easier to read than others:

julia> A, B, C = (rand([1,2,NaN],10) for _ in 1:3);

julia> map(A,B,C) do a,b,c
        !(isnan(a) | isnan(b) | isnan(c))
       end |> findall
3-element Vector{Int64}:
  1
  3
 10

julia> findall(@. !(isnan(A) | isnan(B) | isnan(C)))
3-element Vector{Int64}:
  1
  3
 10

julia> findall(@. !any(isnan, tuple(A,B,C)))
3-element Vector{Int64}:
  1
  3
 10

julia> all(!isnan, hcat(A,B,C); dims=2) |> vec |> findall
3-element Vector{Int64}:
  1
  3
 10

julia> intersect.(.!isnan.(A), .!isnan.(B), .!isnan.(C))
10-element Vector{Vector{Bool}}:
 [1]
 []
 [1]
 []
 []
 []
 []
 [0]
 []
 [1]

julia> intersect(true, true, false)
Bool[]

julia> intersect(true, true, true)
1-element Vector{Bool}:
 1

julia> intersect(false, false, false)
1-element Vector{Bool}:
 0
4 Likes
3

Many thanks for the multiple solutions. They work well :+1: