Build an array with different column lengths

Have one large data set (3333x1 double) and am trying to section out rows based on certain index values. If I have say 10 index pairs (beginning and end row) that I want to section out,each with different lengths how can I build one new nx10 matrix?

I keep getting subscripted assignment mismatches and similar errors when I try things like:

for i = 1:length(indices)
newData(:,i) = olddata(indexstart(i):indexstop(i),:); 
end

Answer 1

[Matt J]

Usually you would store such a result as a cell array,

newData=cell(1,  length(indices));

for i = 1:length(indices)
 newData{i} = olddata(indexstart(i):indexstop(i)); 
end

If it needs to be a numeric array, you need to decide what will go in the missing spots.

[matlabuser12]

SO this will give me one cell array with multiple columns? or multiple cells each with a column in it? How would I achieve the former?

[Matt J]

SO this will give me one cell array with multiple columns? or multiple cells each with a column in it?

Both. A cell array consists of multiple cells. In this case, the cell array that gets created is 1xlength(indices). In other words, it has N=length(indices) columns.

The content of each of the N cells is a numeric vector from olddata. It will be a numeric column vector if your olddata was also a column vector.

[matlabuser12]

I get a 1x10 array with each cell having one column in it. do I need to nest this newdata into another cell array or something?

[Matt J]

Your needs, and what you plan to do with this data later, are completely hidden from me. However, I can't imagine many situations where you couldn't use the 1x10 cell array as is.