Matlab Coder/Code Generation -- Array Size Mismatch

Question

I have the following functioning code (not my own) that I am attempting to put through C++ code generation.

exampleData = double(dataRight(:,2)) < (pointsRight(1,2) + objLength + 0.101);

The right side of the expression fails, however, citing a size mismatch between the left and right side of the less than sign.

Size mismatch (size [:? x 1] ~= size [0 x 0]).

However, matlab's website says that mismatch sizes are, in this case, the typical use case. Further, no specific exceptions or oddities are noted for Matlab Coder, so I'm a bit lost why I'm getting this error.

could you share a dump on pointsRight ? have you actually tried this? as far as i know, mismatch size error is when you want to access non-existing row-column.. — Bagus Tesa
– Bagus Tesa, Commented Aug 28, 2017 at 0:36
[0.051759680404457,0.01,0.0784]. The code runs fine outside of coder's attempt to convert it. Additionally, test = (pointsRight(1,2) + objLength + 0.101) runs fine. It is specifically the introduction of < that causes the error. — Nick Moody
– Nick Moody, Commented Aug 28, 2017 at 0:52
ah i see.. so basically the trouble is (pointsRight(1,2) + objLength + 0.101) scalar (only had a single element), while double(dataRight(:,2)) could have more than a single element (multiple rows).. — Bagus Tesa
– Bagus Tesa, Commented Aug 28, 2017 at 1:09
I'm not sure I understand. Is this a coder-specific restriction? A = [5,4,3]; B= A > 4 works, outside of coder. It yields [1,0,0]. — Nick Moody
– Nick Moody, Commented Aug 28, 2017 at 1:43
hmm, i get your point. it is indeed weird. i think it is the limitation of C++ code generation. — Bagus Tesa
– Bagus Tesa, Commented Aug 28, 2017 at 2:51

Nick Moody · Accepted Answer · 2017-08-28 23:56:27Z

0

Long story short: you must implement a c-style workaround yourself. I believe what caused by issue is the unbounded size of the array, above. I solved this with size(array) and a for loop.

This is the problem code:

dataRight = dataRight(double(dataRight(:,2))>pointsRight(1,2)+0.099 & double(dataRight(:,2))<pointsRight(1,2)+objLength + 0.101,:);

Here is the code that coder accepted:

[widthDataRight,heightDataRight]= size(dataRight)
tempDataRight = zeros(widthDataRight,heightDataRight)
for i = 1:widthDataRight
     val = dataRight(i,2)
     tempDataRight = dataRight(double(val)>pointsRight(1,2)+0.099 & double(val)<pointsRight(1,2)+objLength + 0.101,:);
end

answered Aug 28, 2017 at 23:56

Nick Moody

2631 silver badge12 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Matlab Coder/Code Generation -- Array Size Mismatch

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related