How to delete the duplicate number or using unique in the cell?

Let's see if I have A:

A={[3];[6 8 3];[5];[10 5]}

I am trying to delete the duplicate numbers and found that 'unique' can return the array without the duplication.

Is there any way that I can use unique or delete the duplicate to my cell structure? result like this for example:

A={[3];[6 8];[5];[10]}

Answer 1

[Andrei Bobrov]

A={[3];[6 8 3];[5];[10 5]};

k = repelem((1:numel(A))',cellfun(@numel,A));
B = cell2mat(cellfun(@(x)x(:),A,'un',0));
[a,b] = unique(B,'stable');
out = accumarray(k(b),a,[],@(x){x(:)'});

[Hang Vu]

Thank you! it's perfect.^^

[Hang Vu]

It has a problem:

A={[3];[6 8];[5 10];[10]};
k = repelem((1:numel(A))',cellfun(@numel,A));
B = cell2mat(cellfun(@(x)x(:),A,'un',0));
[a,b] = unique(B,'stable');
A= accumarray(k(b),a,[],@(x){x(:)'})
A{:}

it may return 3x1 cell array in this case.

A =

  3×1 cell array

    {[       3]}
    {1×2 double}
    {1×2 double}
ans =

     3


ans =

     6     8


ans =

     5    10

Could you please help to return the last cell with [] , mean A still has 4x1 cell array

   A={[3];[6 8];[5 10];[]}