How can I use a for loop to index a vector, average the indexed values, and store the averages in another vector when the subsets are of different lengths?

I have column vectors of CO2 flux values that were measured at a 5 minute frequency and correspdoning date/time vectors (Year, Month, Day, Hour, etc).

I am trying to create a for loop to subset each flux value by its corresponding hour, average the flux values for each subsetted hour (the lengths are not the same), and store the averaged values in a vector to represent the average flux for Hour 0-23.

Even just trying to subset I am getting a warning that reads: the variable 'Hour' appears to change size on every loop iteration (within a script). Consider preallocating for speed.

Is there no way to do this using a for loop? Ideally I don't want to use 24 lines of code just to index by Hour = 0-23.

Using MATLAB R2019a

[Star Strider]

We can’t tell from here.

Post your data (or a representative sample), and note the MATLAB release you’re using.

[Michaela Jackson]

Sorry, still new to this. Just posted .mat file and MATLAB release.

[Walter Roberson]

the variable 'Hour' appears to change size on every loop iteration (within a script). Consider preallocating for speed.

Initialize the array Hour to 24 entries before the loop.

[Michaela Jackson]

Perhaps I am just not understanding this at all.

 A = zeros(1,24);        
for i=1:24
  A(i) = find(Hour == i);
end

Now I am getting: Unable to perform assignment because the left and right sides have a different number of elements.

I guess that means I can't store a vector using A(i)?

[Walter Roberson]

No, the problem is the A matrix. You find multiple entries with the same Hour value, but you try to store them in the scalar location A(i) . You could consider using a cell array

A = cell(1,24);
for i=1:24
    A{i} = find(Hour == i);
end

However, considering that you are wanting to calculate average values, you should consider an approach such as

A = zeros(1,24);
for i = 1 : 24
    idx = Hour == i;
    A(i) = mean(flux(idx));
end

Answer 1

[Akira Agata]

You don't need to use for-loop.

How about the following?

load('fluxhour.mat');

[group,hourNum] = findgroups(Hour);
avgFlux = splitapply(@(x) mean(x,'omitnan'),Flux_1,group); % <- since Flux_1 contains NaN values

T = table(hourNum,avgFlux,'VariableNames',{'Hour','avgFlux'});

The output looks like:

>> T

T =

24×2 table

Hour avgFlux

____ ________

0 0.024935

1 0.024158

2 0.02483

3 0.023554

...

[Michaela Jackson]

This is exactly what I wanted to do. Thank you so much!