fill in zeros in a matrix

Hi all, I have a 1000x1 double matrix. It is full of numbers....but on some situations I get multiple 0s.

Whenever I encounter 0's, I want to fill the zeros with the average of the two numbers (the nearest non zero numbers above and below).

For example, if we had 3, 4, 0 ,3 , 2, 2, 0 ,0, 0, 3, 2, 1

we would get 3, 4, 3.5, 3, 2, 2, 2.5, 2.5, 2.5, 3, 2, 1.

The shorter and faster the code, the better....please remember the amount of consecutive zeros vary.

Answer 1

[Jan]

x    = [3, 4, 0 ,3 , 2, 2, 0 ,0, 0, 3, 2, 1];
m    = (x == 0);
x(m) = interp1(find(~m), x(~m), find(m));

Answer 2

[Matt J]

I would recommend replacing the zeros with NaNs and using something like this,

http://www.mathworks.com/matlabcentral/fileexchange/4551-inpaintnans

instead.

[Mate 2u]

Hi, looking for something much shorter and easier to implement, as it was a interview question....and would be nice to have something short which I can remember.