Spliting a cell based on the values of a column

Hi I have a NX2 cell, where the the first column has a data type of the kind int64 and the 2nd column contains strings.

I want to split this cell into smaller cells based on the value of integer in the 1st column. That is if my cell variable is: cell=[1 PB1;3 PB2;1 PBC;2 PB1;1 PBC.......],

I want to create smaller cells out of it; i.e cell_1 which contains all the rows of the original cell which has 1 in their 1st column, cell_2 with 2 in the 1st column and so on . Can this be done without using a for loop, as N can be quite large for me in some cases.

I did it for matrices using logical indexing, but my matrices had only either 1 or2 or 3 or 4 in the 1st column and there I had to specify the value of the elements for carrying out splitting.

Now for the cell array I have values running from 1 upto 290 or more. So,can this be done for cell arrays without exactly specifying what is the value of the element in the column-1 of the cell.

Thanks, Dibakar

Answer 1

[goerk]

Take the data from column 1 and perform the command

 sortInd = unique(column1data);

now you can loop over this values (sortInd) and perform the same as for your matrix solution.

[Dibakar Yadav]

Thanks, for the reply Goerk. This works.

Answer 2

[Stephen23]

There is no need to use any loop:

>> C = {1,'PB1';3,'PB2';1,'PBC';2,'PB1';1,'PBC'}
>> [V,idx] = sort([C{:,1}])
>> D = diff([find([1,diff(V)]),1+numel(V)])
>> out = mat2cell(C(idx,:),D,2)
>> out{:}
ans = 
    [1]    'PB1'
    [1]    'PBC'
    [1]    'PBC'
ans = 
    [2]    'PB1'
ans = 
    [3]    'PB2'

[Dibakar Yadav]

Thanks Stephen.