Comparing uneven sized matrix and delete rows which are not common

Hi guys. I hope this quation finds you well.

For my project i need to compare two uneven matriz row by row. If the rows are not equal beteween column 2 and 5 in a line, delete the line which de index (i+1) is equal to the other matrix index (i). If a rows is deleted, the cycle must restart. Basically i want the column 2 and 5 be exactly equal on both matrices.

For now i have this:

close all;
clc;

D=load('Dir.txt');
L=load('Esq.txt');
D=sortrows(D,6);
L=sortrows(L,6);

for i=1:length(L);
    if D(i , 5) ~= L(i , 5) | D(i , 2) ~= L(i , 2) 
        return
    
    end
end

But i have to manually delete the rows. Is there an option to do this automatically?

I have to delete more than 70000 rows so if i do it manually i'll be done some time next year XD.

Thanks in advance.

P.S: The files are attached

Answer 1

[Ameer Hamza]

Try this code

D = readmatrix('Dir.txt');
E = readmatrix('Esq.txt');

[tf, idx] = ismember(D(:, [2 5]), E(:, [2 5]), 'rows');

D_new = D(tf, :);
E_new = E(idx(idx~=0), :);

Are you looking for something like this?

[David Martins]

Hi Ameer.

Thank you so much for replying to my question.

Infortunately is not. With this code i have a increase of data on de lower sized matrix. The data never gets increase. The final result would be two matrix with the same size and both column 2 and 5 equal.

I have attached to .JPG for a better understanding.

EDIT: I use D=sortrows(D,6); because the 6 column is relataed to time anda i need it to seperate the matrix in other small matrices.

P.S: In the .jpg files, D=D_new and L=E_new on your original code.

[Ameer Hamza]

Can you check the following code. It will not increase the number of rows in any matrix

D = readmatrix('Dir.txt');
E = readmatrix('Esq.txt');

[tf, idx] = ismember(E(:, [2 5]), D(:, [2 5]), 'rows');

D_new = D(idx(idx~=0), :);
E_new = E(tf, :);

Also, if this is also not correct, do you want the rows in columns 2 and 5 of each matrix also to be unique. Creating two small matrices and showing the expected output will help in explaining the problem.

[David Martins]

Still not correct.

let me try to explain better. Imagine the following matrices:

A=[2  5   7  8  6            B=[2  4   6  7  6
   2  4   6  7  6               7  12  4  56 4
   7  12  4  56 4               2  5   7 50  66
   2 78  65 56 56                3 66  54 78 54]
   3 66  54 78 54 
   24 78  65 56 56 ]  
   

The rows 2 from the matrix A is equal to the row 1 form matrix B. Só i need to delete de first row from A. Than we have this:

 A=[2  4   6  7  6            B=[2  4   6  7  6
   23 5   54 55 66               7  12  4  56 4
   3 66  54 78 54                2  5   7 50  66               
   7  12  4  56 4                3 66  54 78 54]
   3 66  54 78 54]

Now we havo to remove de 2nd and 3rd rows of the matrix A.

  A=[2  4   6  7  6            B=[2  4   6  7  6
     7  12  4  56 4               7  12  4  56 4
     3 66  54 78 54]              2  5   7 50  66               
                                  3 66  54 78 54]
  

Finally we ended with the to matrx equal to each other by removing the 3rd row of matrix B

   A=[2  4   6  7  6            B=[2  4   6  7  6
     7  12  4  56 4               7  12  4  56 4
     3 66  54 78 54]              3 66  54 78 54]

I can do this manually but they are to many rows. to delete. I hope that example explained better my problam

EDIT: I forgot to mation that i my case only columns 2 and 5 must be equal on both matrices.

[Ameer Hamza]

The code in my last comment does exactly this

A = [2  5   7  8  6
    2  4   6  7  6
    7  12  4  56 4
    2 78  65 56 56
    3 66  54 78 54
    24 78  65 56 56];

B = [2  4   6  7  6
    7  12  4  56 4
    2  5   7 50  66
    3 66  54 78 54];


[tf, idx] = ismember(B(:, [2 5]), A(:, [2 5]), 'rows');

A_new = A(idx(idx~=0), :);
B_new = B(tf, :);

Result

>> A_new
A_new =
     2     4     6     7     6
     7    12     4    56     4
     3    66    54    78    54
>> B_new
B_new =
     2     4     6     7     6
     7    12     4    56     4
     3    66    54    78    54

The problem arise where there is a repitition in the rows of B. For example, what do you expect the output for following two matrices.

A = [2  5   7  8  6
    2  4   6  7  6
    7  12  4  56 4
    2 78  65 56 56
    3 66  54 78 54
    24 78  65 56 56];

B = [2  4   6  7  6
    7  12  4  56 4
    2  5   7 50  66
    3 66  54 78 54
    11 66  54 70 54];

Note that colums 2 and 5 are same for last two rows of B, but other elements are different.

You may try following code too

D = readmatrix('Dir.txt');
E = readmatrix('Esq.txt');

E = unique(E, 'rows', 'legacy');
[tf, idx] = ismember(E(:, [2 5]), D(:, [2 5]), 'rows');
D_new = D(idx(idx~=0), :);
E_new = E(tf, :);

[David Martins]

It works!!

Thank you so much for your time an effort to enlight me on this matter!

Wish you all the best.

[Ameer Hamza]

I am glad to be of help!