Matlab spdiags function not do what I expect

B=[1 2 3;4 5 6;7 8 9;10 11 12]
d=[-3 0 2];
A=spdiags(B,d,4,7);   %主对角线之下的对角线溢出时候，取最后一个值
B=spdiags(B,d,7,4);   %主对角线之上的对角线溢出时候，取最后一个值
C=spdiags(B,d,4,4);
full(A)
full(B)
full(C)

Why would I get

ans =

     2     0     3     0     0     0     0
     0     5     0     6     0     0     0
     0     0     8     0     9     0     0
    10     0     0    11     0    12     0


ans =

     2     0     9     0
     0     5     0    12
     0     0     8     0
     1     0     0    11
     0     4     0     0
     0     0     7     0
     0     0     0    10


ans =

     0     0     8     0
     0     5     0     0
     0     0     0     0
     2     0     0     0

Why is C so different from the previous two? In terms of having the numbers not shown on main diagnol and also having the numbers arranged from bottom to top？

[Stephen23]

Remember to click the accept button for the answer that best resolves your original question. For example here:

https://www.mathworks.com/matlabcentral/answers/2173595-matlab-sparse-matrix-does-not-come-out-as-i-would-expect

youu have an answer which correctly resolves your question. It would be polite to accept it.

Answer 1

[Stephen23]

"Why is C so different from the previous two?"

Look at your code! On this line you completely overwite B with a completely different variable:

B=spdiags(B,d,7,4);

^ new ^ old

Of course all following lines will refer to that new B, not any old B that might have existed previously.

Solution: do not reuse variable names:

M = [1,2,3;4,5,6;7,8,9;10,11,12]

M = 4×3

1 2 3 4 5 6 7 8 9 10 11 12

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

d = [-3,0,2];
A = spdiags(M,d,4,7);
B = spdiags(M,d,7,4);
C = spdiags(M,d,4,4);
full(A)

ans = 4×7

2 0 3 0 0 0 0 0 5 0 6 0 0 0 0 0 8 0 9 0 0 10 0 0 11 0 12 0

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

full(B)

ans = 7×4

2 0 9 0 0 5 0 12 0 0 8 0 1 0 0 11 0 4 0 0 0 0 7 0 0 0 0 10

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

full(C)

ans = 4×4

2 0 9 0 0 5 0 12 0 0 8 0 1 0 0 11

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

[Zeyuan]

You answered part of my question. My next question would be: why in C, the -3rd diagnol is 1, the first number in col1 of M, while in A, the -3rd diagnol is 10, the last number in col1 of M? In other words, why when spill over, in A, the last number is accepted and in C the first number is accepted？

[Stephen23]

"In other words, why when spill over, in A, the last number is accepted and in C the first number is accepted"

The SPDIAGS documentation explains that behavior here:

https://www.mathworks.com/help/matlab/ref/spdiags.html#mw_6de431d3-8887-4b16-b947-86a610ddea8c

The diagrams in the help neatly illustrate its behavior. Basically you can think of it as slanting those columns at 45 degrees and then selecting the submatrix of the requested size.