Sparse vs Least square solution differ at the edges

Hi, im trying to understand why when im using sparse matrix vs full matrix for solving linear system, i got different values only at the edges.

im using the following code:

l = [12; 6 * ones(98,1); 0];
d = [6; 24*ones(98,1); 6];
u = [0; 6*ones(98,1); 12];
c = spdiags([ l,d,u ],[-1 0 1],100,100);
b = randn(1, 100);

s1 = b/c;
s2 = mldivide(full(c),b')';

figure;plot(s1);hold on;plot(s2);

The sparse part is from spline.m source code.

Answer 1

[Matt J]

Because you forgot to transpose c,

s2 = mldivide(full(c.'),b.').';

[Chen Dadon]

Well thank you for the quick response! It saves me a lot of time!