how set diagonal =1 in matrix tridimensional a(:,:,:)

hi,

how can i set=1 the diagonal of the multidimensional matrix

size(COR)

ans =

8 8 188

[Walter Roberson]

Are you looking for the diagonal on each "page", so all 118 layers the same?

Or are you looking for the "space diagonal" -- (1,1,1), (2,2,2), up to (8,8,8) with everything else 0?

[aldo]

hi.

example: given a matrix a(b,c,d) I want to keep all the values ​​of the matrix but set all the diagonals of the sub-matrixes to =1 (b,c)

[aldo]

[r,c,d]=size(a);

for i=1:d

for x=1:r (it's a square...r=c)

a(x,x,i)=1;

end

end

[Bruno Luong]

So you have for-loop that seems correct and arguably preferable to all answers you'll get. Why asking the question?

[aldo]

because i dont want to use cicle for

I thought it could be solved using Eye /diag without using loops

[aldo]

i use the solution of Walter Roberson:

COR1=COR;

COR2=COR;

tic

[r,c,d]=size(COR1);

for i=1:d

for x=1:r

COR1(x,x,i)=1;

end

end

toc

tic

M = repmat(logical(eye(size(COR2,1),size(COR2,2))),1,1,size(COR2,3));

COR2(M) = 1;

toc

fprintf("equal %d \n",isequal(COR1,COR2));

Elapsed time is 0.001945 seconds.

Elapsed time is 0.000182 seconds.

equal 1

[Bruno Luong]

Here is the timings of three methods

COR = rand(8,8,188);

timeit(@() methodfor(COR)) % Aldo

ans = 4.0301e-05

timeit(@() methodlogical(COR)) % Walter

ans = 2.3801e-05

timeit(@() methodindex(COR)) % Bruno

ans = 1.6800e-05

function COR = methodfor(COR)
[r,c,d]=size(COR);
for i=1:d
     for x=1:r
         COR(x,x,i)=1;
     end    
end    
end

function COR = methodlogical(COR)
M = repmat(logical(eye(size(COR,1),size(COR,2))),1,1,size(COR,3));
COR(M) = 1;
end

function COR = methodindex(COR)
[b,c,d] = size(COR);
COR(1+(b+1)*(0:min(b,c)-1)'+b*c*(0:d-1)) = 1;
end

[aldo]

okk thank

Answer 1

[Geovane Gomes]

Maybe using eye and repmat

COR = repmat(eye(8),1,1,188);

size(COR)

ans = 1×3

8 8 188

[aldo]

is not correct

I want to keep all the values ​​of the matrix (except the diagonal)

[Walter Roberson]

M = repmat(logical(eye(size(COR,1),sie(COR,2))),1,1,size(COR,3));
COR(M) = 1;

[aldo]

Correct! thank you

Answer 2

[Bruno Luong]

% Generate dummy test data
a = 0.01*rand(2,3,4)

a = 
a(:,:,1) =

    0.0058    0.0073    0.0096
    0.0027    0.0011    0.0010


a(:,:,2) =

    0.0081    0.0020    0.0006
    0.0038    0.0068    0.0023


a(:,:,3) =

    0.0052    0.0037    0.0065
    0.0073    0.0024    0.0026


a(:,:,4) =

    0.0057    0.0047    0.0098
    0.0098    0.0082    0.0040

[b,c,d] = size(a);
[I,K] = ndgrid(1:min(b,c),1:d);
a(sub2ind([b,c,d],I,I,K)) = 1;

a

a = 
a(:,:,1) =

    1.0000    0.0073    0.0096
    0.0027    1.0000    0.0010


a(:,:,2) =

    1.0000    0.0020    0.0006
    0.0038    1.0000    0.0023


a(:,:,3) =

    1.0000    0.0037    0.0065
    0.0073    1.0000    0.0026


a(:,:,4) =

    1.0000    0.0047    0.0098
    0.0098    1.0000    0.0040

Answer 3

[Bruno Luong]

% Generate dummy test data

a = 0.01*rand(2,3,4)

a = 
a(:,:,1) =

    0.0023    0.0073    0.0085
    0.0021    0.0024    0.0021


a(:,:,2) =

    0.0073    0.0066    0.0072
    0.0021    0.0099    0.0025


a(:,:,3) =

    0.0035    0.0094    0.0085
    0.0008    0.0069    0.0050


a(:,:,4) =

    0.0070    0.0079    0.0051
    0.0071    0.0029    0.0022

[b,c,d] = size(a);

a(1+(b+1)*(0:min(b,c)-1)'+b*c*(0:d-1)) = 1;
a

a = 
a(:,:,1) =

    1.0000    0.0073    0.0085
    0.0021    1.0000    0.0021


a(:,:,2) =

    1.0000    0.0066    0.0072
    0.0021    1.0000    0.0025


a(:,:,3) =

    1.0000    0.0094    0.0085
    0.0008    1.0000    0.0050


a(:,:,4) =

    1.0000    0.0079    0.0051
    0.0071    1.0000    0.0022