How to crop matrices at the maximal non-NaN values and then center the data

Hi all,

I am new to MATLAB and I have the following problem:

I have matrices something like:

H(1).matrix = [1 2 3
               3 4 5
               3 4 4
               6 NaN 7
               NaN NaN NaN]
           
H(2).matrix = [3 4 5
               5 6 7
               4 4 4
               NaN 3 3
               NaN 4 4
               NaN NaN 7]
H(3).matrix = [3 3 3
               2 1 3
               NaN 1 2]
           

Now, I see that these matrices all have 3 columns but different row number, first is 5x3, second 6x3, third 3x3, but also within columns they have different row length with actual numbers and not NaN. I see that the longest rows length where each columns have non NaN values is 2 (in the third matrix, the first row and the second row with all non NaN values). So now I want to crop all the matrices to have size 2x3. Then I want to center all these matrices.

Something like:

               H(1).matrix = [1 2 3 
                              3 4 5]
               H(2).matrix = [3 4 5
                              5 6 7]
               H(3).matrix = [3 3 3
                              2 1 3]

Then I want to center the data (remove from each element the average of the column it belongs to).

Thank you in advance!

Answer 1

[Andrei Bobrov]

M = struct2cell(H);
n = min(cellfun(@(x)find(any(isnan(x),2),1,'first')-1,M));
for jj = 1:numel(H)
    H(jj).matrix = H(jj).matrix(1:n,:);
end

or without loop 'for-end'

M = struct2cell(H);
n = min(cellfun(@(x)find(any(isnan(x),2),1,'first')-1,M));
C = cellfun(@(x)x(1:n,:),M(:),'un',0);
H = cell2struct(C,'matrix',2);

[Spresi]

I get the error for the first and the second:

Error using cellfun
Non-scalar in Uniform output, at index 1, output 1.
Set 'UniformOutput' to false.
 

[Andrei Bobrov]

Hm. On my desktop:

>> H(1).matrix = [1 2 3
               3 4 5
               3 4 4
               6 NaN 7
               NaN NaN NaN];
           
H(2).matrix = [3 4 5
               5 6 7
               4 4 4
               NaN 3 3
               NaN 4 4
               NaN NaN 7];
H(3).matrix = [3 3 3
               2 1 3
               NaN 1 2]
H = 
  1×3 struct array with fields:
    matrix
>> M = struct2cell(H);
n = min(cellfun(@(x)find(all(~isnan(x),2),1,'last'),M));
for jj = 1:numel(H)
    H(jj).matrix = H(jj).matrix(1:n,:);
end
>> H.matrix
ans =
     1     2     3
     3     4     5
ans =
     3     4     5
     5     6     7
ans =
     3     3     3
     2     1     3
>> H(1).matrix = [1 2 3
               3 4 5
               3 4 4
               6 NaN 7
               NaN NaN NaN];
           
H(2).matrix = [3 4 5
               5 6 7
               4 4 4
               NaN 3 3
               NaN 4 4
               NaN NaN 7];
H(3).matrix = [3 3 3
               2 1 3
               NaN 1 2]
H = 
  1×3 struct array with fields:
    matrix
>> M = struct2cell(H);
n = min(cellfun(@(x)find(all(~isnan(x),2),1,'last'),M));
C = cellfun(@(x)x(1:n,:),M(:),'un',0);
H = cell2struct(C,'matrix',2)
H = 
  3×1 struct array with fields:
    matrix
>> H.matrix
ans =
     1     2     3
     3     4     5
ans =
     3     4     5
     5     6     7
ans =
     3     3     3
     2     1     3
>> 

[Spresi]

I am not sure why it's not working in my case. I have way bigger matrices than what I wrote above, but I think it should work for bigger matrices as well. I keep getting the same error:

Error using cellfun
Non-scalar in Uniform output, at index 1, output 1.
Set 'UniformOutput' to false.

But now, looking at my matrices I see that I also have matrices of the type:

H(4).matrix = [2 3 4
               4 5 6
               7 NaN 8
               9 9 10
               NaN NaN 3
               NaN NaN NaN]
           

Meaning again matrices that have three columns but they may contain NaN in between numbers, and at the end again NaN.

Again I want to crop this one taking into account also previous ones (H(1), H(2), H(3)) to:

               H(4).matrix = [2 3 4
                              4 5 6]
                          

You think this might be causing the problem? Thank you in advance..

[Andrei Bobrov]

A minor fix.

Please attach a small part of your real data as a mat-archive.

[Spresi]

Here is the data. Thank you again!

[Andrei Bobrov]

In your data in all matrices, the value of the 57th column is nan.

[Spresi]

Oh sorry, I did not notice that at all. It works now.. Thank you!!