Averaging Every 5 Elements in a Matrix for Each Column with Nans

I have a matrix of 2241x6914 and I was hoping to average every 5th element of each column such that the final matrix is 448x6914. The issue is that there are locations in which Nans appear in the data and I dont want these to affect this averaging. Ive tried the below code and it generally works but im curious if there is any way for it to not cause a Nan to default the average of the 5 elements to NaN. What I mean by this is that if the values [1,2,3,Nan,4] were being averaged, the value would return Nan as opposed to the 2.5 I would want. Is there any ways around this? I'll also note that im sure this isnt the most efficient way to code this but it just was the most simple way in my eyes. Any help would be appreciated!

Ch3StrainAvg(1,:)=ChtrainData(1,:);
endCondition=size(Ch3StrainData,1)-4
n=2
for j=2:5:endCondition
    Ch3StrainAvg(n,:)=((Ch3StrainData(j,:)+Ch3StrainData(j+1,:)+Ch3StrainData(j+2,:)+Ch3StrainData(j+3,:)+Ch3StrainData(j+4,:))/5);
    n=n+1;
end

Answer 1

[Toder]

For x=[1 2 3 NaN 4], like your short example, you can do this to ignore NaNs in the mean:

mean(x(~isnan(x)))

[Josh]

Is there a way to use indexing with this method so it only takes elements increments of 5 instead of the entire matrix?

[Toder]

You can use x(1:5:end) to get very fifth element of vector x.

[Josh]

Thanks!

[Toder]

Sure thing. Do you mind accepting my answer?

[Josh]

Done!