How can I eliminate unwanted zeros from a matrix?

e.g I want to make this:

1,1,1,1,0,0
1,1,0,0,0,0
1,0,0,0,0,0
1,1,1,1,1,0

into this:

1,1,1,1
1,1
1
1,1,1,1,1
...

Thanks

[Catherine]

This is exactly what I want to do as well! I want to be able to remove all the zeros yet in a matrix form.

I have a matrix:

A =

   557   577   598     0     0     0     0     0     0     0
     0     0     0   558   578   599     0     0     0     0
     0     0     0     0     0     0   558   579     0     0
     0     0     0     0     0     0     0     0   559   579

and I want it to be

A =

   557   577   598
   558   578   599
   NaN   558   579
   NaN   559   579

[Guillaume]

Catherine, please start your own question rather than adding to somebody's else.

[Roger Stafford]

@Catherine: Let A be your original matrix. You can accomplish what you want with:

   T = A~=0;
   n = sum(T,2);
   m = max(n);
   B = nan(size(A,1),m);
   for k = 1:size(A,1)
     B(k,m-n(k)+1:m) = A(k,T(k,:)); % <-- B is the result
   end

Answer 1

[Justin]

It depends on what you are trying to end up with exactly but a good approach would be to set all zero values to NaN. NaNs wont show up in plots and you can use nanmean() or other functions which ignore NaNs to work with your matrix. This will also retain the shape of your matrix.

arr = [1 1 0; 1 0 1; 0 0 1];
arr =
     1     1     0
     1     0     1
     0     0     1
arr(arr==0) = nan
arr =
     1     1   NaN
     1   NaN     1
   NaN   NaN     1

Another option is to use a sparse matrix where the memory only retains information about the index of values.

Sparse

   sparse(arr)
ans =
   (1,1)        1
   (2,1)        1
   (1,2)        1
   (2,3)        1
   (3,3)        1

You could even convert the array into a cell array and use cellfun to replace all the zeros with an empty array. It would be much more difficult to operate on the information then though.

What are you trying to accomplish with the data?

[Mohamad Mossad]

Thanks, setting up zeros to nan is a great idea. 6 years later.

Answer 2

[per isakson]

Hint:

    >> num = [ 1,1,1,1,0,0 ];
    >> num( num == 0 ) = []
    num =
         1     1     1     1

[Justin]

This might not work if there is more than one row of data. For instance:

   arr = [1 1 0; 1 0 1; 0 0 1]
arr =
     1     1     0
     1     0     1
     0     0     1
   arr(arr == 0) = []
arr =
     1     1     1     1     1

[per isakson]

Certainly it will not work with a 2D matrix. However, why should I guess what type of output Mark want for 2D matrices.

[Justin]

Well I assumed he would want a 2D results given the multiple rows shown in his question description. Of course it may not be so but it is important to understand this limitation.

[per isakson]

The correct answer is that for 2D it is not possible. That's because the result will not be a matrix. Why not let Mark find that out himself and return with a clarification.

[Image Analyst]

Or the output could be a cell array where you can have different sized rows in each cell. But I don't know why you'd want to mess with that or even want what he's asking for in the first place. Chances are he won't need what he asked for once we learn what he plans on doing with the result.

[per isakson]

Agree! I know, I should not try to answer questions like this one.