Error using horzcat Dimensions of matrices being concatenated are not consistent.

Please I have been struggling with this problem in my work since weeks. Please I will more than honored for your kind support. The occurs as below:

Error using horzcat

Dimensions of matrices being concatenated are not consistent.

Error in tvf_emd (line 50)

y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)

Please the attached is the code and data used.

[Walter Roberson]

What is the difference between this question and your previous https://www.mathworks.com/matlabcentral/answers/481411-err0r-within-code :?

[Matt J]

Or this one,

https://www.mathworks.com/matlabcentral/answers/480597-error-using-horzcat-dimensions-of-matrices-being-concatenated-are-not-consistent

[Yussif M. Awelisah]

 y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)

   % y=extendsignal(y,num_padding); % padding to deal with boundary effect (match slope) 
   

The first question uses the second definition of y and the second question uses the first definition of y(default in the code).

Please I grateful for your concern. Hope you can help solve this problem.

Answer 1

[Walter Roberson]

temp_x is 500 x 231. num_padding is 250 in that code.

In the expression

        y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)

then y(2:2+num+num_padding-1) is y(2:251) . Using linear indexing of a 500 x 231 matrix, this is the same as y(2:251,1) except for the shape, a part of the first column, except that MATLAB happens to extract it as a row 1 x 250. fliplr() of that 1 x 250 is also 1 x 250.

y(end-num_padding:end-1) is linear indexing, so end has to be treated relative to linear indexing. y(end-num_padding:end-1) is then y(numel(y)-num_padding:numel(y)-1) which is y(500*231-250:500*231-1) which is y(115500-250:500*231-1) which is y(115520:115499) . This is the same as y(250:499, 231) except that MATLAB happens to extract it as a row 1 x 250. fliplr() of that 1 x 250 is also 1 x 250.

Thus the left and right hand side of the [] are both 1 x 250.

The middle, the y, is, as mentioned, 500 x 231. So you are trying to do

   horzcat( 1 x 250 matrix, 500 x 231 matrix, 1 x 250 matrix)
   

However, the number of rows does not match -- 500 rows is not the same thing as one row.

I am unable to offer a correction to your code: I already offered you a suggestion in a previous question (that is, loop over the rows using the row index), but you appear to feel that is not correct for your situation and I have not been able to understand your explanation of why it did not work for you.

[Yussif M. Awelisah]

I am much grateful and humbled. I misrepresented the loop indices when I tried. I hope you help me represent it well in my code again. I think this is the only way.

below is how I misrepresented in the code and the error I got.

 y = temp_x(row_number, :);
  
  flag_stopiter=0;
    for iter=1:100
        for row_number = 1 : size(temp_x,1)
    y = temp_x(row_number, :);
    y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)
  
end

error:

Undefined function or variable "row_number".

Error in tvf_emd (line 43)

y = temp_x(row_number, :);

also IS the second 'for loop' represented correctly?

[Walter Roberson]

Remove the line

  y = temp_x(row_number, :);

that is before the loop.

[Yussif M. Awelisah]

after using this, this occurs,

num_padding = round(size(temp_x,1)*0.5);%  padding number 
    
    y = temp_x;
    flag_stopiter=0;
    for iter=1:100
      for row_number = 1 : size(temp_x,1)
    y = temp_x(row_number, :);
    y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)
   
end

Index exceeds matrix dimensions.

Error in tvf_emd (line 48)

y = [fliplr(y(2:2+num_padding-1)) y fliplr(y(end-num_padding:end-1))]; % padding to deal with boundary effect (symmetric)

[Yussif M. Awelisah]

At this point I have decided to give you the working code and data and the unworking code and data.

tvf_emd(1) is the original code and works perfectly with the data 'one sample'. running both data on tvf_emd(1), may help you understand my error better.