Sort a matrix based on unique values

Hi,

Lets assume I have matrix M:

And I want to fill in matrix O:

I want to sort the final matrix based on the first row in matrix M and O. If I just fill in matrix O without sorting, I get this result:

This is wrong, because colum 3 should be NaN while column 4 should have the values from matrix O, because there is a "match" between the first rows (1004 and 1004).

I hope it was clear. Thanks

[Stephen23]

Use ismember.

If you had given data as text then I would have shown you how, but I can't do anything with data in screenshots.

[Mak]

Hi Stephen,

Thanks for the fast response. It is just random created data. My original data is very large (857 x 19000).

But here is the code for this example:

M = NaN(4,5);

N = [1001 1002 1003 1004 1005];

M = [N;M];

O = rand(2,3);

O2 = [1001 1002 1004];

O = [O2;O];

M(2:4,1:3) = O;

Answer 1

[the cyclist]

I think you can use the ismember function to do what you want. For example,

[tf,loc] = ismember([1001 1002 1004],[1001 1002 1003 1004 1005]);

will result in

loc = [1 2 4]

skipping the 3rd column as you want.

So, you probably want something like

[tf,loc] = ismember(O(1,:),M(1,:));

to find your indices.

[the cyclist]

Side comment: It's generally a bad idea to name a variable "O", because it is so easy to confuse it with a zero. Give your matrices descriptive names. It will make debugging easier.

[Mak]

I have tried this:

if ismember(M(1,:),O(1,:)) == true

M(2:5,1:5) = O

end

But did not work. It's not replacing

[Mak]

This worked. Thank you

M = NaN(4,5);

N = [1001 1002 1003 1004 1005];

M = [N;M];

A = rand(3,3);

B = [1002 1003 1005];

A = [B;A];

idx = find(ismember(M(1,:),A(1,:)))

M(2:5,idx) = A