Interpolate columns in matrix that contain NaN values

Hi I have this code to interpolate data that is set to NaN. I use the fillmissing function as shown below but it does not do the job very well. Could I replace the fillmissing function with interp function?

idx = abs(R_f) <= 0.8;

mat_up(idx,:) = NaN;

mat_up=fillmissing(mat_up,'linear');

[Mathieu NOE]

hello

what do you see as a problem ?

seems to me there is indeed interpolation and NaNs are removed.

load('R_f.mat');
load('mat_up.mat');

idx = abs(R_f) <= 0.8;
mat_up2 = mat_up;
mat_up2(idx,:) = NaN;

mat_up3=fillmissing(mat_up2,'linear');

err = abs(mat_up - mat_up3);
figure(1);imagesc(err);colorbar('vert');
figure(2);imagesc(double(isnan(err)));colorbar('vert');

Answer 1

[Voss]

https://www.mathworks.com/matlabcentral/answers/1836353-interploate-between-rows-in-a-matrix-when-the-corresponding-columns-of-vector-r_f-with-values-rangi#comment_2435393