reshape a 2D to 3D matrix

My data

       ID    X1    X2 
       1     1     2
       1     5     2
       2     2     4
       2     4     7
       3     2     4
       3     4     7
       3     5     11

My goal is to have the submatrix based on the ID. The question is when the rows of each submatrix is not the same, how can i get the following result?

Matrix(:,:,1)=     
     1     1     2
     1     5     2

Matrix(:,:,2)=  
     2     2     4
     2     4     7

Matrix(:,:,3)=  
     3     2     4
     3     4     7
     3     5     11

Answer 1

[Star Strider]

You have to use a cell array, but it is relatively easy:

D = {1     1     2
     1     5     2
     2     2     4
     2     4     7
     3     2     4
     3     4     7
     3     5     11};

Matrix = {D(1:2,:)  D(3:4,:)  D(5:7,:)};
Matrix{:}                                   % Check Result

produces:

ans = 

      [1]    [1]    [2]
      [1]    [5]    [2]

ans = 

      [2]    [2]    [4]
      [2]    [4]    [7]

ans = 

      [3]    [2]    [ 4]
      [3]    [4]    [ 7]
      [3]    [5]    [11]

[Stephen23]

Sure Matrix can be a cell array, but why does D have to be as well? This can still be numeric.

[Yujiao Li]

Thanks a lot. But my goal is to construct a 3D data stucture by loop. so that I can use x(:,:,i) and x(:,:,j) to calculate further. Could you help me on improve my code:

ID.uni=unique(ID); k=length(ID.uni); X=[]; for i=1:k rows_i= D(:,1)==ID.uni(i); % index which rows have the same ID X(:,:,i)=D(rows_i,2:3); end

Thank you so much!

[Star Strider]

My pleasure!

It has to be a cell array, so I would simply do this, taking advantage of the third output of unique, using accumarray as a bin counter, and mat2cell to create the cell array:

D = [1     1     2
     1     5     2
     2     2     4
     2     4     7
     3     2     4
     3     4     7
     3     5     11];

[IDuni,~,ic]=unique(D(:,1));
rows_i = accumarray(ic,1);
X = mat2cell(D, rows_i, 3)

Matrix_1 = X{1}     % Display Resultd
Matrix_2 = X{2}
Matrix_3 = X{3}

producing:

X = 

      [2x3 double]
      [2x3 double]
      [3x3 double]

Matrix_1 =

       1     1     2
       1     5     2

Matrix_2 =

       2     2     4
       2     4     7

Matrix_3 =

       3     2     4
       3     4     7
       3     5    11