Is there an alternative for the ismember() function in a for loop?

I have a program that reads an xlsx file, finds unique rows in the first 3 columns and compares these unique rows to all the rows in the same 3 columns in the initial table that was imported. The goal is to find the indices of the matching rows to find duplicates. There is another column that contain values that I want to compare. So I want to find the duplicate rows and compare their values, and if they're differen, I store them in a new table. The problem is that the imported data is really large, so it takes a very long time for this function to finish what it's doing.

I would be thankful, if you had any suggestions that would help me speed up the process. Here you can see the code that I've described:

[file,path] = uigetfile('*.xlsx');
filename = [path file];


[~,~,raw_file] = xlsread(filename);
tabelle = cell2table(raw_file(2:end, 1:3));
unirow = unique(tabelle, 'rows', 'stable');
daten = cell2table(raw_file(2:end, :));

error_table = table;
idxrow = 1;
datum = datestr(datetime(raw_file(2,2)), 'yyyymmdd');

for ii = 1: size(unirow,1)
    [lia,~] = ismember(tabelle,unirow(ii,:));
    idx = find(lia);
   % idx = find(cellfun(@(c)all(strcmp(unirow(ii,1:3),c)),num2cell(raw_file(2:end,1:3),2)));
    werte = daten(idx, 9);
    if size(unique(werte),1) > 1
        for jj = 1:size(idx, 1)
            error_table(idxrow, :) = daten(idx(jj,1),:);
            idxrow = idxrow + 1;
        end
    end
end

Answer 1

[David Hill]

Have you looked at the unique() function? It might work for you.

a=unique(A,'rows','stable');

[Nigar Musayeva]

in which part do you suggest using it?

The thing is that the table contains multiple duplicates of the same row, and I want to find the indices of duplicates of a certain row at a time and compare their values

[David Hill]

A simplified example would help me understand.

[Nigar Musayeva]

let's say I have a data set like this one

             Var1                           Var2             Var3                  Var4
    ______________________________    ____________    _____________________

    'name1'                            '30.11.2020'    '29.11.2020 16:44:00'      [0,0,0] 
    'name1'                            '30.11.2020'    '29.11.2020 16:44:00'      [1,2,1]  
    'name3'                            '30.11.2020'    '29.11.2020 16:45:00'      [0,0,0]
    'name4'                            '30.11.2020'    '29.11.2020 16:44:00'      [0,0,0]  
    'name1'                            '30.11.2020'    '29.11.2020 16:44:00'      [0,1,0]
    'name3'                            '30.11.2020'    '29.11.2020 16:45:00'      [0,2,0]

so unirows stores theunique rows in columns 1 to 3. Then I use ismember to find the indices of the unique rows (for example row 'name1 30.11.2020 29.11.2020 16:44:00' has indices 1,2 and 5). I use these indices to compare the values in the 4th column.

I hope it's somehow clear:)