Interpolating each row of a matrix with NaN values

I curently have a matrix 'VSFv' with a size of 30x173, with each row a different line, corresponding to my 30 samples. There are negative numbers, creating segmented lines in some samples. I am trying to interpolate those segmented lines to create one continuous line, replacing the negative numbers in each row in a for loop. I am able to do it with one row (row 18) with the corresponding code:

inter = VSFv(18,:);
inter(inter<0) = NaN;
interp = ~isnan(inter);
ave = cumsum(interp-diff([1,interp])/2);
final = interp1(1:nnz(interp),inter(interp),ave);

The for loop that I am trying to work out is:

for j=1:1:size(T,1)/2
    inter(j,:) = VSFv(j,:);
    inter(inter<0) = NaN;
    interp(j,:) = ~isnan(inter(j,:));
    ave(j,:) = cumsum(interp(j,:)-diff([1,interp(j,:)])/2);
    F(j,:) = interp1((interp(j,:)),inter(interp),ave(j,:));
end

Where j=30, inter is my matrix with NaNs instead of negative numbers, and ave is my query vector. I get the error

"Error using interp1>reshapeAndSortXandV (line 424)

X and V must be of the same length."

I also tried using 'griddedInterpolant' code too, but had no luck. Can anyone help?

Answer 1

[Star Strider]

I am not certain that I understand the problem, however if the intent is to interpolate the NaN values, see if the fillmissing function (introduced in R2016b) will do what you want.

[Kacey Lange]

Just tried it and it worked! Used this:

F(j,:) = fillmissing(inter(j,:),'movmedian',10);

Which fixed my problem. Thank you!

[Star Strider]

As always, my pleasure!