splitting matrix based on a value in one column

Hi All,

I have a very large matrix part of which looks like this:

   7  0
  10  0
  16  0
  19  1
  26  1
  29  1
  3  1
  5  0
  69  0
  36  0
  78  0

as you can see, the second column is either 0 or 1. I want to split this large matrix into sub matrices such that each sub matrix represent only the parts where the second column is 0. How do we do this? Thanks.

Answer 1

[Stephen23]

Hard-coded indexing:

>> M = [7,0;10,0;16,0;19,1;26,1;29,1;3,1;5,0;69,0;36,0;78,0];
>> M(M(:,2)==0,:)
ans =
     7     0
    10     0
    16     0
     5     0
    69     0
    36     0
    78     0
>> M(M(:,2)==1,:)
ans =
    19     1
    26     1
    29     1
     3     1

General solution:

>> [~,~,X] = unique(M(:,2));
>> C = accumarray(X,1:size(M,1),[],@(r){M(r,:)});
>> C{1}
ans =
     7     0
    10     0
    16     0
    36     0
    69     0
    78     0
     5     0
>> C{2}
ans =
    26     1
    19     1
    29     1
     3     1

[Thomas Booij]

Thanks a lot!

[Aleem Akhtar]

That was great solution. You saved day for me.

[Sandeep Karthikeyan]

Thanks a lot for the answer.

[Joshua]

Great solution! How would you do this for the transpose of M? I.e. if I had the array M' and wanted to sort the columns into a cell matrix of arrays?

[Chris Stillo]

Thank god someone knows what they are doing!

[Pol Medir]

How would you do this if the sample of data was too large to input manually?

[Stephen23]

@Pol Medir: import the data:

https://www.mathworks.com/help/matlab/data-import-and-export.html

Answer 2

[123456]

How could I do the same thing but divide cell array by strings from one row? Let's say, that the array looks like that:

t2   1    2
t2   2    3
t1   3    4
t2   4    5
t1   5    6
t1   6    7

I would like to recieve two arrays at the end:

1    2
2    3
4    5

and

3    4
5    6
6    7

[Stephen23]

There are several solutions. One way is to input the first column into unique, and use its third output with my answer.