Using methods across multiple class objects of the same class?

Hello everyone

In short: What is the best way to implement a method that uses the properties of multiple objects of the same class?

In long: I have multiple sets of data from experiments that I would like to analyze. Experiments happen at certain operating conditions with minor changes across individual experiment runs, with a run meaning to take N measurements at M locations over P seconds. So the data can be split according to operating point (OP), and each operating point according to run (R). Analysis with Matlab should happen on both levels.

My current approach is to collect and evaluate the R-data in a custom class, with the relevant pseudo-code:

%in ..\Project\@RunData\RunData.m
classdef RunData
    properties
        % various properties like: data, information about experiment, e.g., sampling rate and signal length
        fooProp1
        fooProp2
        % ...
    end
    
    methods
        function obj=RunData(fooFile)
            % Constructor
            if nargin==0
                %some properties=NaN
            end
            % initialize some other properties
            [fooProp1, fooProp2]=load(fooFile);
        end
        
        function testPrint(obj)
            fprintf('Value of foo is %s\n',exampleProp1)
        end
        
        % ...
            
    end
end

Now here was the main problem: I want to combine foo1 of multiple objects of class RunData. Let's say, as an example, that I want to get the average of foo1 across all objects.

Thus, my first question: Is it possible to define a function, e.g. "calculateAverage", within the class methods that would evaluate the property combined across each object of that class?

--> Depending on how the question above is answered, the rest below might be completely irrelevant.

Tinkering around with the code and gathering resources from various sources online, I then tried creating an "enveloping class" on OP-level that would hold all the individual runs, on which I could run "array-functions", e.g., get the average of fooProp1 across all runs, with the pseudo-code:

%in ..\Project\@OperatingPoint\OperatingPoint.m
classdef OperatingPoint
    properties
        Run RunData % An array comprised of RunData elements
        fooProp3
        fooProp4
        % ... 
    end
    
    methods
        function obj=OperatingPoint(fooFolder)
            % Constructor that creates an array entry for each data file
            listOfFiles=getDataFilesInFolder(fooFolder)
            for i=1:NumberOfDataFiles
                obj.Run(i)=RunData(listOfFiles(i))
            end
        end
    end
end

Onto the second question: Is there a more elegant approach to the problem than the one I described above?

And finally, as my third question: Is it possible to define the set/get methods so that the call:

>> MyOperatingPoint.Run.fooProp1

would result in an array of fooProp1 of all runs that are in the variable MyOperatingPoint instead of Matlab iterating over all runs and displaying fooProp1 of each individual run?

For anyone that has read this far, thank you very much for your patience!

[Elst]

I just realized the (admittedly trivial) solution to my third question that works to some extent. Simply use square brackets to contain the results, i.e.:

>> [MyOperatingPoint.Run.fooProp1]

Works for fooProp1 being a double, but produces one long mess if fooProp1 is a character array.

[Stephen23]

Generally:

https://www.mathworks.com/help/matlab/matlab_prog/comma-separated-lists.html

https://www.mathworks.com/matlabcentral/answers/1656435-tutorial-comma-separated-lists-and-how-to-use-them

[Elst]

Very interesting read, thank you for making me aware of this!

I will try to apply this knowledge to methods in the "enveloping class" that contains the individual runs.

Answer 1

[Image Analyst]

If the results are strings, try

{MyOperatingPoint.Run.fooProp1}

or

vertcat(MyOperatingPoint.Run.fooProp1)

[Elst]

Good call, thanks for your reply!

First option works on strings and character arrays and gives a cell array as output. But just as a quick note: If the results are strings and already defined as such (in the classdef), square brackets still work for those that want to avoid cell arrays.

Second option only works for character arrays of equal length, otherwise Matlab throws "Dimension of matrices being concatenated are not consistent" (which makes sense).

Thus, the third question is answered. I am looking forward to what people come up with for the other 2 points.