Try this:
fid = fopen('temp.txt','rt');
C = textscan(fid,'%[^\n]','HeaderLines',3);
fclose(fid);
rgx = '^(\d+)\s+(\d+)\s+((\s\d+)+)\s+(\d+(\.\d+))\s*$';
spl = regexp(C{1},rgx,'tokens','once');
spl = vertcat(spl{:});
out = cellfun(@(s)sscanf(s, '%f').', spl, 'UniformOutput',false)
This generates a cell array out, the third "column" contains vectors of values, the other columns are scalar numeric:
>> out =
[1] [3] [3x1 double] [0.9100]
[2] [5] [5x1 double] [0.8200]
[3] [2] [2x1 double] [0.6500]
>> out{1,3}
ans =
1 4 5
Because you did not provide us with a test file I had to create my own, which is available here:
