Fill gaps with zeros in a non-consecutive time series

Hi There,

I understand this might be a simple problem, but I have spent a lot of time on it and can't seem to quite figure it out. I have a series of data A: A = [1,1,1,2,3,4,7,7,8,9,9,11,12,16] and want it to look like B: B = [1,1,1,2,3,4,NaN,NaN,7,7,8,9,NaN,11,12,NaN,NaN,NaN,16]

by finding the gaps (e.g. between 4 and 7 or 9 and 11) and fill those with NaNs. While there are a number of elegant solutions for filling gaps in time series, the issue here is that sometimes the numbers are repeating (i.e. as in [1,1,1] or [7,7]) and the gaps are not always 1.

I appreciate any suggestions!

Answer 1

[Greg Dionne]

This should get you started:

function y = pennyanswer(x)

validateattributes(x,{'numeric'},{'row','finite','integer','nondecreasing'})

% build destination index vector
d = diff(x);
d(d==0) = 1;
ivec = cumsum([1 d]);

% build destination vector, y, (pre-populate with NaN).
y = nan(1,ivec(end));

% assign x to proper location in y
y(ivec) = x;

>> pennyanswer([1 1 1 2 3 4 7 7 8 9 9 11 12 16])

ans =

Columns 1 through 18

       1     1     1     2     3     4   NaN   NaN     7     7     8     9     9   NaN    11    12   NaN   NaN

Columns 19 through 20

   NaN    16

[penny]

wow this is great - you make it look easy! I didn't think of using the cumsum function...but that is a good idea. Thanks!!