# Remove NaN Entries in a Dataset

12 views (last 30 days)
Allen Hammack on 20 Jan 2023
Commented: Allen Hammack on 30 Jan 2023
I have a data set with multiple variables (var1, var2, var3). Any of the variables can have a NaN values. If one value for one of the variables is a NaN, I'd like to remove that instance for each variable. I have the following code that does this, but is very slow:
count = 1;
for j = 1:length(var1)
if isnan(var1(j)) == 1 || ...
isnan(var2(j)) == 1 || ...
isnan(var3(j)) == 1
else
var1_cull(count,1) = var1(j);
var2_cull(count,1) = var2(j);
var3_cull(count,1) = var3(j);
count = count + 1;
end
end
How can I modify this routine to have the speed up the data removal?
I've tried using the find() function, but I haven't been able to get it to work when checking to see if each variable is a nan.
Thank you!
Allen Hammack on 30 Jan 2023
I have incorporated your suggestion and everything is working perfectly! Thank you!

Vilém Frynta on 20 Jan 2023
Edited: Vilém Frynta on 20 Jan 2023
You do not need to go through every element of vector with for loop.
You can use isnan on the vector, which will give you logical vector (0s and 1s), which you can then use on all the other vectors. You can also combine logical vectors together.
I will try my best to demonstrate:
% Random vars for demonstration
var1 = [1, 2, 3, NaN, 5, 6];
var2 = [9, NaN, 7, 6, 5, 4];
% 1 = NaN values, we will invert this later (~)
idx1 = isnan(var1);
idx2 = isnan(var2)
idx2 = 1×6 logical array
0 1 0 0 0 0
% Combine logical vectors
idx = idx1 | idx2
idx = 1×6 logical array
0 1 0 1 0 0
% Apply logical vector to your variables
var1(~idx)
ans = 1×4
1 3 5 6
var2(~idx)
ans = 1×4
9 7 5 4

### Categories

Find more on Logical in Help Center and File Exchange

R2021b

### Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!