How to stop unwanted rows being added when constructing a matrix in a for loop
3 views (last 30 days)
Show older comments
Hi there,
just a simple question I think.
I am trying to construct a matrix in a quick way using a for loop. Because the values for diagonals are the same I can using indexing like (i,i) and (i+1,i) etc. However, apart from the first diagonal I input (i,i), the second diagonal (i,i+1) adds an extra row and column to my original matrix.
Here is my code:
clear, clc, close all
mat = zeros(7,7)
n = size(mat,1)
for i = 1:n
mat(i,i) = 6
mat(i,i+1) = -4
end
If you run the code, the diagonal containing -4 adds on a extra row and column at the end; making a 7x7 matrix to an 8x8. This is very frustraing and I do not want this!
Can I ask what its the trick to getting around this, please?
Many thanks
Scott
1 Comment
Accepted Answer
Star Strider
on 5 Sep 2024
The reason is that there are the same number of ‘6’ and ‘-4’ being created in the loop. There needs to be one less ‘-4’.
clear, clc, close all
mat = zeros(7,7);
n = size(mat,1);
for i = 1:n
mat(i,i) = 6;
mat(i,i+1) = -4;
end
mat
N = 7;
mat2 = diag(ones(1,N)*6); % Use 'diag'
mat2 = + mat2 + diag(ones(1,N-1)*-4, 1)
.
0 Comments
More Answers (3)
ScottB
on 5 Sep 2024
mat = zeros(7,7)
n = size(mat,1)
for i = 1:n
mat(i,i) = 6
if i ==7
else
mat(i,i+1) = -4
end
end
mat =
6 -4 0 0 0 0 0
0 6 -4 0 0 0 0
0 0 6 -4 0 0 0
0 0 0 6 -4 0 0
0 0 0 0 6 -4 0
0 0 0 0 0 6 -4
0 0 0 0 0 0 6
1 Comment
Voss
on 5 Sep 2024
Or
mat = zeros(7,7);
n = size(mat,1);
for i = 1:n
mat(i,i) = 6;
if i ~= n
mat(i,i+1) = -4;
end
end
mat
Steven Lord
on 5 Sep 2024
Since you're creating matrices with diagonal bands, consider using the diag or spdiags functions.
n = 7;
mainDiagonal = diag(6*ones(n, 1))
upperDiagonal = diag(-4*ones(n-1, 1), 1)
A = mainDiagonal + upperDiagonal
Note that when I created upperDiagonal I used a vector of all -4 values that had length one shorter than the vector of 6 values I used to create mainDiagonal. This made it so upperDiagonal was n-by-n rather than (n+1)-by-(n+1) as it would have been had I used -4*ones(n, 1).
B = diag(-4*ones(n, 1), 1)
For bands below the diagonal, use a negative value as the second input.
lowerDiagonal = diag(-8*ones(n-1, 1), -1)
Here are the numbers for which band contains each element of the matrix.
bandNumbers = toeplitz(0:-1:-(n-1), 0:(n-1))
For spdiags you don't have to create each band individually, though the syntax is a little bit more complicated. The command below puts 6 on the 0 (main) diagonal of S and -4 on the +1 (upper) diagonal, and makes S an n-by-n matrix.
S = spdiags([6 -4], [0 1], n, n)
Since S is sparse it's displayed slightly differently. Convert it to full and it looks like A above.
F = full(S)
0 Comments
See Also
Categories
Find more on Data Type Identification in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!