Error ' Assignment has more non-singleton rhs dimensions than non-singleton..'

I want to find the location of maximum 'EC' in the first direction. Here is the code I used to find so.

 for ii=1:size(EC,2);
   ss(1,ii,:)=max(EC(:,ii,:),[],1);
   loc(1,ii,:) = find(EC(:,ii,:) == ss(1,ii,:));
 end

but an error ' Assignment has more non-singleton rhs dimensions than non-singleton subscripts' keeps coming. Can anyone tell me what to do with it?

[Walter Roberson]

Does it occur on the max() or on the find() ?

You have not indicated whether the array is 2D or more. If it is 2D then EC(:,ii,:) would be the same as EC(:,ii) and then max() of that would be a scalar and could be stored in a scalar ss(1,ii,:). Then in the find(), EC(:,ii,:) would be EC(:,ii) a vector, and that could be tested against the scalar ss(1,ii) giving a vector result for the "==". find() on that vector result could result in a scalar (if the max was unique) or could result in a vector (if there were multiple copies of the maximum.)

If, though, EC has more than 2 dimensions, then max(EC(:,ii,:),[],1) would return an array whose first two dimension were and whose third dimension is the product of the remaining dimensions after the second. e.g., if EC was 2 x 3 x 4 x 5 then the max() would be 1 x 1 x (4*5). As this would be a vector, you would be able to store it into ss(1,ii,:). But then in the find(), EC(:,ii,:) would be size(EC,1) by 1 by the product of the dimensions after the second, and you would not be able to "==" that against the vector ss(1,ii,:) unless the first dimension of EC happened to be 1.

[Ede gerlderlands]

It occurs on find(). EC is 3D.

[Ede gerlderlands]

is there any alternative to finding the 'loc' then?

[Walter Roberson]

If you are trying to find the maximum of each column along with the associated indices, use the two-output version of max()

[ss(1,ii,:), loc(1,ii,:)] = max(EC(:,ii,:),[],1);

However you need to decide which index you want returned if there are multiple copies of the maximum in any given column.

[Ede gerlderlands]

Thank you very much...