"Subscripted assignment dimension mismatch." error

Hi,

I'm trying to write this loop here and it's returning the error mentioned above. The matrix dimensions are:

 287x16150  location
 287x16150  cum_return
 287x16150  index
 1093x21    ME_Breakpoints 
 300x16150  Monthly_return

for k = 1:287 %for each month
location(k,:)=find(Cum_Return(k,:) > ME_Breakpoints(13+k,3));
[~,index(k,:)] = sort(Cum_Return(k,:), 'descend');
sorted_index(k,:) = intersect(index(k,:), location, 'stable');
portfolio_return= mean(Monthly_Return(:,index(k,1))-mean(Monthly_Return(:,index(k,1))))
end

Thanks!

[Stephen23]

@Ming Au: please show us the complete error message. This means all of the red text.

[Ming Au]

Hi Stephen, the title of this post was the whole error message:

"Subscripted assignment dimension mismatch."

[the cyclist]

This is difficult to diagnose, because that error could have come from a couple different lines. Could you upload the actual data in a MAT file?

Answer 1

[Image Analyst]

Ming, you could have done what Stephen asked. Anyway, find() can return any number of elements - it varies. But when you do this:

location(k,:)=find(Cum_Return(k,:) > ME_Breakpoints(13+k,3));

you're trying to set all 16150 columns in row k. But what if find returns only 50 elements? It can't stuff 50 elements into 16150 elements. You have to stuff only as many as you have, which is 50. I've fixed that but the error occurs later also, so see if you can fix that yourself.

location = rand(287,16150);
Cum_Return = rand(287,16150);
index = rand(287,16150);
ME_Breakpoints = rand(1093,21);
Monthly_Return = rand(300, 16150);

for k = 1:287 %for each month
  indexes = find(Cum_Return(k,:) > ME_Breakpoints(13+k,3));
  numIndexes = length(indexes);
  location(k,1:numIndexes) = indexes;
  [~,index(k,:)] = sort(Cum_Return(k,:), 'descend');
  sorted_index(k,:) = intersect(index(k,:), location(k,1:numIndexes), 'stable');
  portfolio_return= mean(Monthly_Return(:,index(k,1))-mean(Monthly_Return(:,index(k,1))))
end