unsure whether to resample or interp(1)?

hi! i have some data that needs to be resampled? i think. The original data is varying in row length dependant on the test. ie. test 1 could have 1600 rows, test two could have 1100 and test 3 could have 3000 rows.

should i be re-sampling or interp?

currently i am resampling but i am not getting the same number of rows for each walk. I am sampling data at a frequency of 100Hz. i ideally would like them to be 1000 rows each.

i have tried

resample(new_location,1,1); %and 
resample(new_location,FS,1,10) %WHERE FS IS 100Hz
% % 

i have attached an excel file of my data where the columns are the following respectively.

foot contact #, time, left/right foot, x co-ordinate, y co-ordinate.

this is the code i have so far.

zero = COP(:,2) == 0;
COP(zero,:) = [];
COP = COP;

finalrow = length(COP);
pts = findchangepts(COP(:,2), 'Statistic', 'linear', 'MinThreshold',50);
pts(end+1) = finalrow; %adds another point to array (final)
next = 1;

k_footfall = COP(:,1);
time = COP(:,2);
left_right_f = COP(:,3);
x = COP(:,4);
y = COP(:,5);
bigarray = {};

% this section of code finds the  row location of the new walk and flips
% the corresponding rows so that the rows all start from the the same
% origin. 

for i=1:length(pts);
    if mod(i,2) == 0
        new_location = [k_footfall(next:pts(i,:)), time(next:pts(i,:)),left_right_f(next:pts(i,:)), x(next:pts(i,:)), y(next:pts(i,:)) ];
        new_sample(:,1) = i;
    else
        new_location = [k_footfall(next:pts(i,:)), time(next:pts(i,:)),left_right_f(next:pts(i,:)), 350- (x(next:pts(i,:))), y(next:pts(i,:))];
        new_sample = resample(new_location,1,1);
        new_sample(:,1) = i;
    end
    next = pts(i);
    bigarray{i} = new_location; 
    re_sample{i} = new_sample;
   
end

%vertically concatenating the data
COP_location = vertcat(bigarray{:})
resample_x = vertcat(re_sample{:});

%finds the unique values. ie. how many data points for each walk
[C2,iac,icb] = unique(resample_x(:,1));
a_counts_walk = accumarray(icb,1);
value_counts_walk = [C2, a_counts_walk];

[Mathieu NOE]

hello Caroline

could you clarify if files length are different because have different sample rates and / or different time duration ?

so the target is only to reduce to same length(1000 rows each) independantly from original sampling frequency ?

tx

[caroline bourn]

file lengths are different because they have a different time duration.

i want to get them having the same number of rows by interpolating the data, i think?

[Mathieu NOE]

hello

IMHO, you have to interpolate the data to a common time vector

[caroline bourn]

thank you for that! is there a function you would suggest? :)

[Mathieu NOE]

hello Caroline

I would try with interp1

all the best