Search multiple matrices for values at common index positions
9 views (last 30 days)
Show older comments
I am trying to analyse several tens of 90x90 matrices and I want to find a summary matrix that only shows a logical 1 if an element has any non zero value that is common to all the matrices at the same index position. In the example below the desired result is to show that there is a value in common between all matrices in position [2,2] and [4,4]. I do not wish to show the values themselves or compare their values. I've tried working with ismember , find , intersect , a==b for just two matrices but nothing seems to do the trick. I can't seem to find a similar query in the archives either. Does anyone know of a suitable function or do I need to write an element by element search routine across all the matrices . All matrices are double and all identically sized .
a =
1 0 0 0 0
0 3 0 0 0
0 0 5 0 0
0 0 0 7 8
>> d = [ 0 0 0 12 0; 0 3 0 4 0; 0 0 0 0 0 ; 34 0 0 9 0]
d =
0 0 0 12 0
0 3 0 4 0
0 0 0 0 0
34 0 0 9 0
>> e = [ 0 0 0 0 0; 0 15 0 0 0; 0 0 0 2 0 ; 0 0 0 5 0]
e =
0 0 0 0 0
0 15 0 0 0
0 0 0 2 0
0 0 0 5 0
>> Result=[ 0 0 0 0 0; 0 1 0 0 0; 0 0 0 0 0 ; 0 0 0 1 0]
Result =
0 0 0 0 0
0 1 0 0 0
0 0 0 0 0
0 0 0 1 0
6 Comments
Stephen23
on 24 Jun 2015
Edited: Stephen23
on 24 Jun 2015
Learn to use the dimensions of arrays and MATLAB will fly...
MATLAB is a high-level language, and it works best when data is kept together as much as possible. Avoid nesting, multiple variables and loops (unless you really have to). Then you can use vectorized code to really get the best out of MATLAB: neat and fast operations.
Good luck and have fun!
Accepted Answer
James Tursa
on 23 Jun 2015
Edited: James Tursa
on 23 Jun 2015
If you are just looking for when there are non-zero entries in the same index in all matrices, e.g.,
Result = a & d & e;
To process "several tens" of these matrices, you may want to write a function. How the function is written will depend on how you have your matrices stored (individual variables, members of a cell array or struct, etc). E.g., if they are individual variables, then a function like this might be a good start:
function g = allnz(varargin)
n = nargin;
if( n )
g = logical(varargin{1});
for k=2:n
g = g & varargin{k};
end
else
g = [];
end
Then you could call it like this:
Result = allnz(a,d,e);
2 Comments
Guillaume
on 24 Jun 2015
In my opinion, a better implementation (no loop) of your allnz would be:
function g = allnz(varargin)
dimplus1 = ndims(varargin{1}) + 1;
g = all(cat(dimplus1, varargin{:}), dimplus1);
end
More Answers (4)
Guillaume
on 23 Jun 2015
Edited: Guillaume
on 24 Jun 2015
As given in the example:
result = all(cat(3, a, d, e), 3)
Note that if you tens of these matrices, rather than putting them in individual variables, you're better off putting them in a cell array:
allmatrices = {a, b, c, d, e, f, ...};
In which case the above is simply:
result = all(cat(3, allmatrices{:}), 3);
Or even better since they all have the same size, just stack them all up in the third dimension as I have done as a first step, so:
allmatrices = cat(3, a, b, c, d, e, f, ...);
In which case the code is simply:
result = all(allmatrices, 3);
Azzi Abdelmalek
on 23 Jun 2015
Edited: Azzi Abdelmalek
on 23 Jun 2015
a=[ 1 0 0 0 0 ; 0 3 0 0 0 ; 0 0 5 0 0 ; 0 0 0 7 8]
d=[ 0 0 0 12 0; 0 3 0 4 0; 0 0 0 0 0 ; 34 0 0 9 0]
e=[ 0 0 0 0 0; 0 15 0 0 0; 0 0 0 2 0 ; 0 0 0 5 0]
aa=~~a
dd=~~d
ee=~~e
aa&dd&ee
Martin Brown
on 24 Jun 2015
As above, reformat your code to put the matrices into a single 3D matrix, something like:
A(:,:,1) = [1 0; 0 1];
A(:,:,2) = [2 0; 0 0];
find(sum(A(:,:,1:end)~=0,3)==2);
0 Comments
See Also
Categories
Find more on Matrix Indexing in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!