Interpolation of nan values

I have a column with some Nans, and I'd like to interpolate the data to put numbers instead of the missing values.

The nans are in column 2.

Column 3 is the flag, so I'm looking for values that are number 4(flagged as missing/nan) and changing them to a 1(flagged as changed interpolated data), after the change is done.

This is what I tried:

numRows = size(database,1);

    for i=1:numRows
       if database(i,3) == 4;
        database(i,2) = interp1(database(i,2),'linear')
        database(i,3) = 1;
       end

[Stephen23]

Do not use i or j for variable names, as these are both names of the inbuilt imaginary unit.

Answer 1

[Stephen23]

Here is an example of how to achieve something like what you need:

>> B(:,2) = [1,2,NaN,4,NaN,6,7,NaN,9];
>> B(:,3) = [0,0,  4,0,  4,0 0,  4,0]
B =
     0     1     0
     0     2     0
     0   NaN     4
     0     4     0
     0   NaN     4
     0     6     0
     0     7     0
     0   NaN     4
     0     9     0
>> idx = B(:,3)==4;
>> B(idx,2) = interp1(find(~idx), B(~idx,2), find(idx));
>> B(idx,3) = 1
B =
     0     1     0
     0     2     0
     0     3     1
     0     4     0
     0     5     1
     0     6     0
     0     7     0
     0     8     1
     0     9     0

Note that this completely avoids using any loop at all (i.e. fully vectorized code), and that it uses a syntax that taken from the interp documentation.

Answer 2

[Walter Roberson]

Get John D'Errico's inpaint_nans