Averaging sections of a given column

Suppose I have a column vector A=[1 3 -999 4 9 8 -999 3 5]; I need to get average of each 3 sections. The problem is, if any section contains -999, I need to avoid it and get the average of rest of the numbers. For an example, for the 1st three numbers, average has to be (1+3)/2.

Please help me with that. Thanks.

[MSP]

clear all
close all

A=[1 3 -999 4 9 8 -999 3 5]
index=find(A==-999)
A(index)=0;
sectionsize=3;
k=1;
for i=1:3:size(A,2)-2
   out=A(i:i+2)
   nzeros=numel(find(out==0));
   output(k)=sum(out)/(sectionsize-nzeros);
   k=k+1;
end
output

[Thishan Dharshana Karandana Gamalathge]

Thanks. But using this code, if there were three -999 somewhere in the middle, that section will be replaced by the avg of next three. At the same time, if there are three -999 at the end, it will be not shown at all. In those cases, let's put NaN so that I know which sections had three consecutive -999. Consider the following to help me with coding.

A=[1 3 -999 -999 -999 -999 4 9 8 -999 -999 -999]; The correct answer I am expecting is B=[2 NaN 7 NaN]

Thanks.

[Image Analyst]

That seems inadvisable. Why do you want nans in there instead of the very simple approach I showed you below?????

MSP, that should have been in the answer section below since you intended it as an answer rather than a comment (like asking for more information).

Answer 1

[Image Analyst]

If you have the Image Processing Toolbox you can use regionprops():

A=[1 3 -999 4 9 8 -999 3 5]
props=regionprops(A~=-999, A, 'MeanIntensity');
sectionMeans = [props.MeanIntensity]

You get:

sectionMeans =
     2     7     4