Re-write repeated values of a matrix using intermediate increased values

Hi all, I have some problems to rewrite a matrix with the values that I need. The following code do this:

• Step0 = a matrix is created
• Step1 = sort my matrix with ascend order
• Step2 = create a mask with zeros
• Step3 = here mask is rewrited with zeros just in repeated values positions
• Step4 = create a matrix with repeated values changed by zeros.
• Step5 = need to rewrite zero values with intermediate increased values (help me here !!)

%%step0: My matrix 
A = [ 1 1 1 1 2 2 3 4 5 7 5; 
       3 3 6 5 4 1 5 7 8 9 1 ] ;
%%step1: sort my matrix ascend order
Asorted = sort(A,2); 
%%step2: create a mask
Mask = zeros( size(Asorted,1),size(Asorted,2)); 
%%step3: Now Mask is rewrited with zeros in repeated values positons
for i = 1: size(Asorted,1)
    [ C,ia,ic ] = unique( Asorted(i,:) );
    Mask( i, ia ) = 1;   
end
%%step4: vlaues are sorted with repeat val eliminated
AsortedWihtoutRepeat = Asorted.*Mask
% step5: Now we need to rewrite zero values with intermediate increased values
% RewritedMatrix = ??

Step4 results:

AsortedWihtoutRepeat =

     1     0     0     0     2     0     3     4     5     0     7
     1     0     3     0     4     5     0     6     7     8     9

Now I need to rewrite zeros by intermediate values, for example:

AsortedWihtoutRepeat(1,2) = value been value bigger than 1 and lower than 2. AsortedWihtoutRepeat(1,3) = value been value bigger than 1 and lower than 2.

And I need that position (1,2) and (1,3) will be increased values.

Step5 example:

RewritedMatrix =

     1     1.1     1.5     1.7     2     2.5     3       4     5     6     7
     1       2       3     3.5     4       5     5.5     6     7     8     9

Answer 1

[Guillaume]

Another option is to use fillmissing with the 'linear' option (which under the hood does call interp1 similar to Stephen's answer). You can simplify your code to:

A = [ 1 1 1 1 2 2 3 4 5 7 5; 
     3 3 6 5 4 1 5 7 8 9 1 ] ;

%step 1:
Asorted = sort(A, 2);
%step 2, 3 and 4 combined:
Asorted([false(size(A, 1), 1), diff(Asorted, [], 2) == 0]) = NaN;
%step 5:
Afinal = fillmissing(Asorted', 'linear')'  %has to transpose since fillmissing works on columns

Note that if you worked on columns rather than rows, the whole code would be simpler:

A = [ 1 1 1 1 2 2 3 4 5 7 5; 
     3 3 6 5 4 1 5 7 8 9 1 ]';  %transposed A

Asorted = sort(A);
Asorted([false(1, size(A, 2)); diff(A) == 0]) = NaN;
Afinal = fillmissing(Asorted, 'linear')

[Cesar Antonio Lopez Segura]

Thank you very much

Answer 2

[Stephen23]

Just use interp1 on each row of your matrix:

>> vec = [1,0,0,0,2,0,3,4,5,0,7]; % one row of the matrix
>> idx = vec~=0;
>> out = interp1(find(idx),vec(idx),1:numel(vec))
out =
   1.0000   1.2500   1.5000   1.7500   2.0000   2.5000   3.0000   4.0000   5.0000   6.0000   7.0000

Put that in a for loop and you can apply it to each matrix row.

[Cesar Antonio Lopez Segura]

Thank you for your answer. I find some problems when matrix is like this:

A = [ 1 1 1 1 2 2 3 5 5 5 5; 3 3 6 5 4 1 5 7 8 9 1 ] ;

[Stephen23]

@Cesar Antonio Lopez Segura: that is a matrix, not a vector.

What are the problems? Did you use a loop, as I wrote in my answer?

[Cesar Antonio Lopez Segura]

Here the code with the loop

clc; clear all
% step0: My matrix 
A = [ 1 1 1 1 2 2 3 5 5 5 5; 
     3 3 6 5 4 1 5 7 8 9 1 ] ;
% step1: sort my matrix ascend order
Asorted = sort(A,2); 
% step2: create a mask
Mask = zeros( size(Asorted,1),size(Asorted,2)); 
% step3: Now Mask is rewrited with zeros in repeated values positons
for i = 1: size(Asorted,1)
    [ C,ia,ic ] = unique( Asorted(i,:) );
    Mask( i, ia ) = 1;   
end
% step4: vlaues are sorted with repeat val eliminated
AsortedWihtoutRepeat = Asorted.*Mask;
% step5: Now we need to rewrite zero values with intermediate increased values
for i = 1: size(AsortedWihtoutRepeat,1)
vec = AsortedWihtoutRepeat( i,: ); % one row of the matrix
idx = vec~=0;
RewritedMatrix( i,: ) = interp1(find(idx),vec(idx),1:numel( vec ) );
end
RewritedMatrix

The command window response

RewritedMatrix =

    1.0000    1.2500    1.5000    1.7500    2.0000    2.5000    3.0000    5.0000       NaN       NaN       NaN
    1.0000    2.0000    3.0000    3.5000    4.0000    5.0000    5.5000    6.0000    7.0000    8.0000    9.0000

[Stephen23]

@Cesar Antonio Lopez Segura: somehow you have changed your algorithm. In your original question you gave this example matrix:

AsortedWihtoutRepeat =
   1   0   0   0   2   0   3   4   5   0   7
   1   0   3   0   4   5   0   6   7   8   9

But when I actually run your code I get this matrix:

AsortedWihtoutRepeat =
   0   0   0   1   0   2   3   4   0   5   7
   0   1   0   3   4   0   5   6   7   8   9

Given your different input arrangement, there is no reason to expect the same output.

[Cesar Antonio Lopez Segura]

Thank you very much.

I want to explain that my real matrix has 1600 rows and 32 colum.

For this reason I was looking for a general solution that works with my problem using a simplify example. The aim is manipulate the gen of a lot of chromosomes in a genetic algorithm and not a number response of interp1 is a problem in my case.

Best regards Cesar