Reproducing elements from one vector to another.

Hi!

I have a problem that I can't wrap my head around. Two data sets, one with a matrix with values (depth x time), and a time vector, and the other with a vector of values, along with a time vector. The time step is three hours. The first set however, is irregular, with some time steps being repeated (due to a higher temporal resolution), looks like this:

Now, how should one do to get the elements from the second data set replicated in the same way as the first? I.e. like using repelem but with irregular increments. Appreciate any help, thanks!

Johan

[Jan]

This sounds like a job for ismember - both outputs are needed. So please post both data sets in a form, which can be used by copy&paste, such that we can test out suggestions. A screenshot of one of the data sets is not useful.

[Johan Edholm]

Thanks for the tips, will include everything next time. Did manage to solve it with unique, find, isfinite, double, and sum. By first separating the unique values in time_543, and making a new vector, with all the time steps from the first to the last, with 3 hours increments, as well as a empty vector to fill with how many times each time step is repeated, or zero.

aa = unique(time_543);
cc = datenum(2011,12,09,09,00,00):0.125:datenum(2012,03,06,06,00,00);
bb = zeros(1,length(cc));

for i=1:length(time_543)
    I(i) = find(cc==aa(i),1);
end
    
dd(I) = aa;

for i=1:704
    bb(i) = sum(double(isfinite(find(time_543==dd(i)))));
end

After this, I just used repelem with the other data set and bb. Hope this is clear enough for anyone having the same problem.

[Jan]

This is not really meaningful:

bb(i) = sum(double(isfinite(find(time_543==dd(i)))));

find() replies the indices of non-zero elements of its input. Indices are finite in every case, so the isfinite() is not useful here. sum() operates on logcial arrays directly, so the casting to double() is not required. You can get the same result by:

bb(i) = sum(time_543 == dd(i));

The approach

cc = datenum(2011,12,09,09,00,00):0.125:datenum(2012,03,06,06,00,00);

is fragil due to rounding errors, see https://www.mathworks.com/matlabcentral/answers/57444-faq-why-is-0-3-0-2-0-1-not-equal-to-zero

I'm still not sure what you want to achieve actually.