How to interpolate only NaN values with one hole?

I have a large matrix (23x6939) with many NaN values. I need to interpolate over any gaps that are only one column wide, but leave any gaps that are two or more NaN values wide.

For example,

x=[1 2 4 4 NaN 6 9 8; 12 13 NaN NaN 20 21 24] becomes [1 2 4 4 5 6 9 8; 12 13 NaN NaN 20 21 24].

How can I distinguish between gaps that are only one value and gaps that are more than one value before applying interp1?

Answer 1

[Dyuman Joshi]

fillmissing is a better fit here -

%Note that the data you have provided is not valid
%I have added another element to make it compatible for concatenation
%Sample data
x = [1 2 4 4 NaN 6 9 8; 12 13 NaN NaN 20 21 24 25; 45 NaN NaN NaN 66:69; 3 5 7 11 13 17 19 23]

x = 4x8

1 2 4 4 NaN 6 9 8 12 13 NaN NaN 20 21 24 25 45 NaN NaN NaN 66 67 68 69 3 5 7 11 13 17 19 23

%Fill the gaps less than or equal to MaxGap with linear interpolation,
%along the 2nd dimension
y = fillmissing(x, 'linear', 2, 'MaxGap', 2)

y = 4x8

1 2 4 4 5 6 9 8 12 13 NaN NaN 20 21 24 25 45 NaN NaN NaN 66 67 68 69 3 5 7 11 13 17 19 23

Refer to the MaxGap Name-value Argument section of fillmissing() for information regarding how this feature is defined.

[Dyuman Joshi]

Any updates, @Gavin Plume?

[Gavin Plume]

Worked like a charm. Thank so much for your help!

[Dyuman Joshi]

You're welcome!