Index exceeds the number of array elements in for loop

Dears, hope you are doing fantastic.

I have a script that I use to find wavelength and wavenumber by uploading text files and define the values for x an y.

It goes ok for one file, however; with mutliple files I have an "Index exceeds the number of array elements" error

A help will be grestly appreciated, thanks :)

The script goes this way:

[file_list, path_n] = uigetfile('.txt', 'Multiselect', 'on');
filesSorted = natsortfiles(file_list);

if iscell(filesSorted) == 0;
    filesSorted = (filesSorted);
end

for i = 1:length(filesSorted);
    filename = filesSorted{i};
    data = load([path_n filename]);
    
    % Define x and y from the uploaded files
    x =  data (:,1);
    y =  data (:,2);
    
    % Calculate Wavelength and Wavenumber K from the uploaded files

[maxvaly,idx] = max(y);

maxvalx = x(idx);

[idx,idx]=findpeaks(y);

Wavelength(i)=x(idx(2))-x(idx(1))

Wavenumber(i)=(2*pi)/Wavelength(i)

end

Answer 1

[Walter Roberson]

    Wavelength(i)=x(idx(2))-x(idx(1))

Your code is assuming that findpeaks() will always be able to find at least two peaks. You should not assume that. You should be checking length() of the result.

[Hussein Kokash]

Any idea how to edit it Walter?

Where can I introduce length() to the script?

TIA

[Walter Roberson]

[peakValues, idx] = findpeaks(y);
if length(idx) >= 2
   Wavelength(i)=x(idx(2))-x(idx(1));
else
   Wavelength(i) = nan;
   warning('Not enough peaks for file: "%s', filenaname);
end

[Hussein Kokash]

To begin with, thank you so much for the contribution.

I just have one thing to add.

If you have a look at the image below, I choose two files, the second one shows wavelength as NaN, but looking at the graph, the is a wave shape, thus two peaks, what could be the cause to that?

Note: the values of y for the case of NaN here are to the power of -11 or -12 so they are pretty close to 0 but since we are taking the index, I don't think that matters, right?

Thanks again Walter

[Hussein Kokash]

Never mind, you were absolutely right, the shape of the wave was irregular and that is why I had a NaN output, it has to be fitted into a function first and then get the wavelength properly.

Thank you and god bless!

[Walter Roberson]

sorry, I do not know much about the Algorithm. You could experiment with findpeaks of (say) 1000 times the input signal.

Answer 2

[Sulaymon Eshkabilov]

The problem is with this 3rd party fcn file: filesSorted = natsortfiles(file_list);

Note that the variable file_list contains th being imported file name, e.g.: 'ABC123.txt'

So, the err is occuring on: filesSorted = natsortfiles(file_list);

and then filename = filesSorted{i}; is not giving the following file name as you've anticipated.

[Hussein Kokash]

Thanks Sulaymon, What can be done here to solve it and keep the pattern of importing files intact?

[Stephen23]

"What can be done here to solve it and keep the pattern of importing files intact?"

Nothing is required to "solve" it, because NATSORTFILES is certainly not the cause of that error.

Lets consider the possible configurations of UIGETFILE's first output:

• user selects multiple files: NATSORTFILES sorts and and returns a cell array of char vectors,
• user selects one file: NATSORTFILES returns the one char vector,
• user selects zero files: NATSORTFILES throws an error due to the invalid input data class (but this error is moot, because your code cannot anyway handle this case: you can decide if you want to handle this case yourself and add the appropriate code).

However there is an obvious bug in your code, because you used parentheses instead of curly braces here:

filesSorted = (filesSorted);
%             ^           ^ these do nothing at all

It should be:

if ~iscell(filesSorted)
    filesSorted = {filesSorted};
end %             ^           ^ you need curly braces!

or simply replace the entire IF-statement with this:

filesSorted = cellstr(filesSorted);

That bug will cause an error to be thrown later in your code (but is totally unrelated to your NATSORTFILES useage, or the error you gave in your question, or anything written in this incorrect answer).

Answer 3

[Sulaymon Eshkabilov]

Easy solution is to get all file names to be read (to be imported in *.txt). That can be done with:

Finfo = dir('*.txt');
Fnames = {Finfo.name};

for i = 1:length(Fnames);
    filename = Fnames{i};
    ...
        
end

[Hussein Kokash]

I replaced what I have with what you suggested, unfortunately still the same error of:

Index exceeds the number of array elements (0).

Error in test (line 40)

Wavelength(i)=x(idx(2))-x(idx(1))

Answer 4

[Image Analyst]

[idx,idx]=findpeaks(y);

is not right. It should be

[peakValues, indexesOfPeaks] = findpeaks(y);

[Hussein Kokash]

It gives the same error? should I change "Wavelength(i)=x(idx(2))-x(idx(1))" as well?

[Image Analyst]

Of course, there is no badly-named, ambiguous idx variable anymore. I fixed that by separating them out into two nicely named variables.

What is Wavelength? Is it the delta X value?

[Hussein Kokash]

Oh I see.

Wavelength basically is the coressponding x values for the highest two consecutive y values.

For instance, it is 20-10 in the example below, 20 and 10 are the corresponding values for the highest two consecutive y values, so how can it be expressed with your script?

Thanks for your help, much appreciated

[Image Analyst]

The you can do

% Get the x values of all the peaks
xPeaks = x(indexesOfPeaks);
% Find distance between adjacent peaks.
peakSpacings = diff(xPeaks);
% Find the mean peak-to-peak spacing.
Wavelength = mean(peakSpacings)

[Hussein Kokash]

Thank you for your valuable input, appreciate it, will try it for sure.