MATLAB Answers

Unusual NaN appearing from a loop

1 view (last 30 days)
Nishant Karve
Nishant Karve on 3 Dec 2018
Commented: OCDER on 3 Dec 2018
Hi there.
I've been trying to run a loop to calculate the quartiles for each of the four columns in a dataset. The loop is as given below:
sized = size(data);
for i = 1:4
Minimum(i,1) = min(data(:,i));
Quart2(i,1) = median(data(:,i));
if rem(sized(1),2) == 0
Quart1(i,1) = median(data(data(:,1) < Quart2(i,1)));
Quart3(i,1) = median(data(Quart2(i,1) < data(:,i)));
Quart1(i,1) = median(Minimum(i,1): Quart2(i,1));
Quart3(i,1) = median((Quart2(i,1)):(max(data(:,i))));
Maximum(i,1) = max(data(:,i));
I'm generating a table of these results after the loop and I get something like this once I run the script:
The problem with this is that I don't know why the 'NaN' values appear here. When I run the line of code evaluating 'Quart1' independently in the command window for this very dataset, it gives me the correct value of 0. I'd be really grateful if someone could help me with this.

Answers (1)

OCDER on 3 Dec 2018
It seems that you are taking the median of an empty number
data = [1 2 3];
Quart2 = 4;
Quart1 = median(data(Quart2 < data));
Quart1 =
Double check that there is no empty values. Otherwise, replace NaN's with 0's.
OCDER on 3 Dec 2018
When you do the logical index:
Quart2(i,1) < data(:,i)
It's possible that NO DATA agrees with this condition. So when you do this:
data(Quart2(i,1) < data(:,i)),
you get NO DATA, or an empty array, []. That's equivalent to doing this:
Quart1 = median([]) = NaN.
To fix, do this:
Quart1(isnan(Quart1)) = 0;

Sign in to comment.

Community Treasure Hunt

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

Start Hunting!