Bypassing the sampling limits of resample function of p*q<2^31

So I have 3 arrays which I want to resample into equal length. I am using the code

C1=resample(c1,target_length,length(c1));
C2=resample(c2,target_length,length(c2));
C3=resample(c3,target_length,length(c3));

where target_length is 70000. The matrix sizes of c1,c2, and c3 are 1x72901, 1x72392, and 1x723532. As expected it gives an error due to the "The product of the downsample factor Q and the upsample factor P must be less than 2^31." I cannot use interp1 as I need unique points too. Is there a way around this or to bypass this restriction for this problem?

[Mathieu NOE]

hello

I don't understand this : I cannot use interp1 as I need unique points too

[Mathieu NOE]

maybe you could share some data where you think you cannot use interp1

[r]

Data is too large so I will just write it down. I got it from a data set and it has about 72k values per cell in a 15x3 cell dataset. The interp1 is throwing an error for these values saying that these are not unique. This isn't the whole array just a few values.

voltage= [4.18794059753418,4.18097162246704,4.18064355850220,4.18045473098755,4.18021249771118...]
    Capacity=[0,3.36861362059911e-05,0.000101058623856968,0.000168430987331602,0.000235803160402510,...
    0.000303175490763452,0.000370547398924828,0.000437919091847208]

[Meet]

Could you share the complete code along with interp1 function you used?

[r]

%original length of c1,c2,c3,v1,v2,v3>70000
%shortening it to fit each others
% Define target length
target_length = 50000; % Desired length for resampling

% equally spaced voltage arrays as max and min are the same;
Vmax=cellfun(@max,v2);
Vmin=cellfun(@max,v2);

V1=linspace(Vmax(1,1),Vmin(1,1),target_length);
V2=linspace(Vmax(1,2),Vmin(1,2),target_length);
V3=linspace(Vmax(1,3),Vmin(1,3),target_length);

C1=interp1(c1,v1,V1);
C2=interp1(c2,v2,V2);
C3=interp1(c3,v3,V3);

[r]

@Meet this is the dropbox link

https://www.dropbox.com/scl/fo/3ss0age6ggfcm67okldhw/AGRF4ccKKYI4z6RW6b2-Uzg/diagnostic_tests?rlkey=pcvgs0ndk6uawo9ho0cbluxrz&subfolder_nav_tracking=1&dl=0

You can go to the processed mat folder and run the data analysis file.

https://www.sciencedirect.com/science/article/pii/S2352340922002062

This is the paper explaning the files. I am currently interested in only the capacity test so I extracted the values for all the 15 batteries and stored them in .mat files.

Answer 1

[Abhas]

Hi @r,

The issue arises because the "interp1" function in MATLAB requires the "x" array to have unique values, and your voltage data (the x input) may not satisfy this condition. To bypass the limitations of "interp1" or "resample", you can use a custom approach to handle duplicate or non-unique "x" values by following the below steps:

• Identify and remove duplicate "x" values by averaging the corresponding "y" (capacity) values. This step ensures "interp1" works correctly.
• After handling duplicates, use "interp1" to interpolate the data to the desired "target_length".
• If "c1", "c2", or "c3" has too large or too small lengths, adjust them accordingly to avoid errors during resampling.

Here's a sample MATLAB code to achieve the same:

target_length = 70000;

c1 = rand(1, 72901);
c2 = rand(1, 72392);
c3 = rand(1, 723532);

% Example voltage arrays (replace with actual voltage data)
V1 = linspace(min(c1), max(c1), length(c1));
V2 = linspace(min(c2), max(c2), length(c2));
V3 = linspace(min(c3), max(c3), length(c3));

% Resample each dataset
C1 = handle_duplicates_and_resample(V1, c1, target_length);
C2 = handle_duplicates_and_resample(V2, c2, target_length);
C3 = handle_duplicates_and_resample(V3, c3, target_length);

% Display results
disp(size(C1));
disp(size(C2));
disp(size(C3));

% Function to handle duplicates and resample
function resampled_data = handle_duplicates_and_resample(x, y, target_length)
    % Ensure unique x values
    [x_unique, ~, idx] = unique(x, 'stable'); % 'stable' keeps original order
    y_unique = accumarray(idx, y, [], @mean); % Average y values for duplicates
    
    % Resample using interp1
    x_resampled = linspace(min(x_unique), max(x_unique), target_length);
    resampled_data = interp1(x_unique, y_unique, x_resampled, 'linear');
end

You may refer to the below MathWorks documentation links to know more about the same:

1. https://www.mathworks.com/help/matlab/ref/double.unique.html
2. https://www.mathworks.com/help/matlab/ref/accumarray.html
3. https://www.mathworks.com/help/matlab/ref/double.interp1.html

I hope this helps!