Extract the value according to the number of rows

Hello everyone, I need to extract or reshape the array value (cixd) so that the number of value is based on the num of rows(picture A) and store at the next column. Here shows some illustration. The extraction continues until it goes to 10895 that follows the value of picture A. I tried used reshape but the error says I cannot change the num of the element.

combineResultDisengged=vertcat(FeaturesDis{:});
s{k,:}=size(FeaturesDis{k},1) %extract value rows only
end
[cidxD,cmeansE] = kmeans(combineResultDisengged,500);
R=reshape(cidxD,s{:,1})
end

[Adam]

You are trying to reshape the entire cidxD array every time and since all you seem to want to do is extract segments of different sizes from it you should just extract these into a cell array. You don't need reshape if they will all remain as column matrices.

[Siti Suhaila]

I've done simple extraction as below but the result only shows the correct one on the first column only. I need it to correctly display on the next column as well. Would you please advise? I know I shouldn't put 1:s{i}. But when I put s{i} itself it cannot gives any result at all.

   M_extract{i} = cidxD(1:s{i},:);

[Walter Roberson]

The second case, the 197 that follows the 185: should that 197 be from the beginning of the vector, or should it be the next 197 elements after the 185 already "used up" ?

[Walter Roberson]

"I need it to correctly display on the next column as well."

That is not possible for a numeric array, as you will have columns of different lengths.

The exception would be if you were to use a "fill value" that got put in to all of the places where there was no value because the column was shorter.

[Siti Suhaila]

The second case, 197 should be the next 197 elements after the 185 already "used up".

[Siti Suhaila]

    R = reshape(cidxE, 10, []);%extract each 10 rows

However, I don't want to restrict it to 10 rows only. I need to extract the rows based on the value of picture A and continues to the next column for the second rows (197) and so on. Did I clearly explain it?

Answer 1

[Walter Roberson]

Let the size information be in A. Let the array to be split be B.

total_requested = sum(A);
if total_requested > length(B)
  error('A requests more entries than there are in B');
end
B_split = mat2cell(B(1:total_requested), A, 1);

Now B_split is a cell array with column vectors of varying lengths. However you cannot arrange them beside each other until you define how to handle the "holes"

[Siti Suhaila]

May I know what will be 'sum(A)'? Because it will give value the end of the rows in picture A not the value of each row.

[Walter Roberson]

Since you are using consecutive blocks from the second matrix, it is good practice to check first if the total amount of entries that would be requested exceeds the actual size of the second matrix. sum(A) is the total number of entries requested.

Here I do not assume that the 132 lengths in the first matrix add up to exactly the length of the second matrix: I check first that they are not longer than what is available, and then I only use the number requested (which accounts for the possibility that the total is shorter than the number of rows available.)

[Siti Suhaila]

I managed to get the correct value of the first rows and the matrix display the number accordingly, however, it started with the first and duplicate it to the next column.

[Walter Roberson]

Are you referring to the fact that M_extract{1,1}(185) is 197, and 197 also happens to be the second length to use according to picture A ?

We do not have your data to test with.