Why isn't this file loading method working?

I've had success in the past piecing together proper file strings name to call upon the right file name. Here I'm trying to compare saved data sets and plot their data series one at a time, since all data sets share the same variable names.

I figured I could just loop through each file name and plot their respective series independent of one another simply by using the 'hold on' syntax to keep previous data sets since all files of interest I want to compare all have the same variable names. That makes sense right? I would just grab one variable of a data set, plot it and hold on, and rewrite over the variable name for the next data set of interest.

I've attached a code screen shot for reference. The two files/datasets I'm working with right now can be seen in the current folder workspace. Like I've said, I've used this technique in the past but cannot recognize why the data sets are not being loaded.

Any advice would be appreciated!

[Stephen23]

Rather than loading directly into the workspace, it is much more robust to load into an output variable:

S = load(..);

[Kieran Fung]

So I was able to make the code run after calling on the "right" part of load(myfilename).

I ended up making the following edits seen below:

I didn't realize I had to call the string of interest through the cell brackets. But can you elaborate on how your suggestion is more robust than whats being implemented below? I am always interested in learning how to be a better coder and I do not really recognize your suggestion.

Thanks!

Kieran

for i = 1:length(theta)
    myfilename(i) = A + num2str(theta(i)) + B
    load(myfilename{i})
    plot(timestep, ConcentrationAverage, 'LineWidth', 3)
    hold on
end

[DGM]

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval?s_tid=ta_ans_results#answer_274740

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval?s_tid=ta_ans_results#answer_236127

The rest of this thread is relevant, but this is a key starting point regarding load().

[Kieran Fung]

thank you!

[Kieran Fung]

Hey DGM and Stephen,

I recognized both of your precense in the thread and wanted to ask a followup question after reviewing that link. Although I didn't understand everything that was being discussed, proposed, etc. - WRT the suggestion of:

s = load(...)

To my understanding, this is more helpful for the solver because the outputted 's' will remain constant throughout my loop, even though the actual loading file will not? Why is this beneficial? I'm not fully connecting the dots because the loaded file name will be chaning periodically for 'N' amount of files, even though 'S' will be constant.

Thanks!

- dynamic struct newbie

[DGM]

Someone can offer a more technically thorough explanation than I can, but if you do something like

load('mybigfilefullofstuff.mat');
y = x*2;
z = mean(y);

Prior to running, or at parse time, how does Matlab know anything about where x came from? It can't know how many other things are going to show up in the workspace until it actually loads the file at run time. It defeats a lot of the guidance that the code tools could give the person writing the code, and it leaves the possibility for all sorts of hard to detect/predict shenanigans to occur. What if one of the variables in that .mat file is called mean?

If you load into a struct:

S = load('mybigfilefullofstuff.mat');
y = S.x*2;
z = mean(y);

Matlab knows ahead of time that y is a function of a thing that came from S which came from load(). It knows that mean() isn't something that came from S.

Answer 1

[DGM]

myfilename is a string containing the filename

load(myfilename) 

passes the string "theta_stuff_acoustics.mat" to load() (and should work if it's a valid filename)

load('myfilename')

passes the char vector 'myfilename' to load(), but there probably isn't any file called 'myfilename'