How to use "if statement" for different matrix?

i dont know how to explain my question,

lets say i have 4 matrix

a = [1 3 5; 2 4 7];

b = [1 2 5; 2 4 6];

c = [1 4 5; 4 4 7];

d = [1 3 1; 2 4 1];

i want use like

for i=1:4

if a, b and d then

x=2+n

AK= a*n*x

BK= b*n*x

DK=b*n*x

else (only c)

x=7+n

CK=c*n*x

end

end

kind of like that?

hope you understand. Thanks

[Steven Lord]

What do you mean by "if a, b and d then"? If a, b, and d what?

• If a, b, and d exist
• If a, b, and d are the same size
• If a, b, and d all contain the same values
• If a, b, and d all contain nonzero values

Or do you want to perform the same steps for a, b, and d and to perform a different set of steps for c?

[sky walker]

yes, i want to exclude c, because c have different step than other

Answer 1

[Stephen23]

The MATLAB approach is to use arrays and indexing, e.g. using one simple cell array:

V = {[1,3,5;2,4,7];[1,2,5;2,4,6];[1,4,5;4,4,7];[1,3,1;2,4,1]};
W = V;
n = 1;
for k = 1:numel(V)
    if k<4
        x = 2+n;
    else
        x = 7+n;
    end
    W{k} = V{k}*n*x;
end
W{:}

ans = 2×3

3 9 15 6 12 21

ans = 2×3

3 6 15 6 12 18

ans = 2×3

3 12 15 12 12 21

ans = 2×3

8 24 8 16 32 8

Or by just using numeric arrays (which is probably what you should be doing, compare how much simpler and more efficient your MATLAB code would be):

A = cat(3,V{:});
U = cat(3,2,2,2,7);
B = A.*n.*(U+n)

B = 
B(:,:,1) =

     3     9    15
     6    12    21


B(:,:,2) =

     3     6    15
     6    12    18


B(:,:,3) =

     3    12    15
    12    12    21


B(:,:,4) =

     8    24     8
    16    32     8

[sky walker]

thanks for your answer Stephen, but my point is to exclude matrix c. because c have different step than other.

and i cant use your way, because i have lot matrix, not just 4.

thanks

[Stephen23]

"but my point is to exclude matrix c. because c have different step than other."

Did you look at my code? The fourth matrix uses a different step, just like you want.

".. cant use your way, because i have lot matrix, not just 4."

If you have to process lots of matrices, then your approach will require writing lots of code. More code means more bugs. More code obfuscates the intent. More code is harder to maintain.

MATLAB was designed around using matrices and arrays to store data. You can solve your problem very neatly and efficiently using matrices and arrays, just like I showed in my answer (and like every experienced MATLAB user would do). This would be simpler and more efficient (most likely in terms of run-time and also your own time).

If you have lots of separate matrices in the workspace then you really need to redesign your data.

[sky walker]

Thanks Stephen, I'm still new to matlab and sorry for my poor english.

I don't know how to express my question.

My code is actually simple,but just generates a lot of new matrices.

And the new matrix needs to be processed again.

and some of them have different step.

I've seen your script, but what i need to be different is matrix c, not d.

[Stephen23]

"My code is actually simple,but just generates a lot of new matrices."

Hopefully in one cell array or structure, otherwise you really have a major design problem.

"I've seen your script, but what i need to be different is matrix c, not d."

My answer does not have matrix c nor d.

Instead of having lots of separate variables (bad data design, difficult to extend to lots of matrices) I simply placed all of your matrices into one cell array (much better data design, works automatically for any number of matrices) or into one ND-array (ditto). Both of these data designs makes it trivial and efficient to loop over your matrices (unlike your complex approach) and also to apply different steps to whichever of those matrices that you want (in fact, which ones does not even need to be hardcoded).

In my answer the fourth matrix has a different step, just as you requested.

You can make any matrix/matrices that you want have different step sizes.

As far as I can tell, my does what you need**, just simpler, more efficient, and more versatile than your approach (and also trivially extends to any number of matrices, unlike your approach).

** or atleast something very similar to it, so similar that you should not have any problems tweaking it.

[sky walker]

ok thanks for your help stephen