problem with eval notation

I have a question about eval in this script
function sort=sortstructure(strucvector, field)
if isfield(strucvector,field)
for i=1:length(strucvector)-1
low=i
for j=i+1:length(strucvector)
if eval(['strucvector(' num2str(j) ' ). ' field])<...
eval(['strucvector(' num2str(low) ' ). ' field])
low=j
end
end
temp=strucvector(i)
strucvector(i)=strucvector(low)
strucvector(low)=temp
end
else disp('wrong field')
end
sort=strucvector
Why cab't I write the part with eval like this:
if eval(['strucvector(num2str(j)).''field'])<...
eval(['strucvector(num2str(low)).''field'])
I know eval is often considered unhandy but it is part of curriculum

1 Comment

The eval is only there because the curriculum is dated to before MATLAB handled dynamic field names.

Sign in to comment.

 Accepted Answer

I don't see any reason for eval above. If you get rid of it, it should still run. If you're required by law of the professor to use it, have it do something useless and not dangerous at the end.
E.g.
%the rest of the code
eval('disp(''we did it!'')')

3 Comments

I tried to erase eval and it stilled worked like this:
if (['strucvector(' num2str(j) ' ). ' field])<...
(['strucvector(' num2str(low) ' ). ' field])
Then i erased a space after the first period and it did not work:
if (['strucvector(' num2str(j) ' ).' field])<...
(['strucvector(' num2str(low) ' ). ' field])
I just get trapped in not understanding where the spaces and '' should be. Is there any system?
I believe in MATALB x.y, where x is smaller than I care to go back to figure out its actual value, the eval was required.
@Tor, it is not really working when you remove the eval. You are in fact doing a comparison of two strings. It gives you an error (I am guessing something about matrix dimensions), when you delete a space from the j side and not a space in the low side.

Sign in to comment.

More Answers (2)

You cannot do
eval(['strucvector(num2str(j)).''field'])
because
['strucvector(num2str(j)).''field']
expands to a string of strucvector(num2str(j)).'field (notice the odd '). EVen if you eliminate the odd ', you won't get the correct answer.
strucvector(num2str(j)).field
will essentially do
strucvector(double(num2str(j))).field
which is not:
strucvector(j).xxx
where xxx is whatever string is contained in field.
That is why you cannot do what you want and the original version works.
Current versions of MATLAB all for
strucvector(j).(field)
Hi Tor, Not sure why you want to use eval() here, but:
test = struct('data',randn(10,1));
for nn = 1:9
if (eval('test.data(nn) < test.data(nn+1)'))
disp('hi'); end
end

Categories

Find more on Operators and Elementary Operations in Help Center and File Exchange

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!