Randomly Appearing Zeros in Data Output

Hello!

I'm still working on some code I have asked about in previous questions. All seems to be going well except in my output I keep getting zeros unexpectedly and I can't seem to figure out why

The code opens data, uses accumarray to place the data into vectors and then writes the output of the vectors into a matrix. For some reason I seem to get random rows of zero in this matrix - and I cant seem to find out why

The code is below:

% Cycle through files rownum = 1;

for f = 1: numberOfFiles resFileName = fullfile(res_files,baseFileNames(f).name) Name = baseFileNames(f).name rawData = importdata(resFileName,' ');

   timeData = rawData.data(:,1);
   ampData = rawData.data(:,2);
   PIDget = Name(3:5)
   PID = str2num(PIDget)
   Task1.PID{f} = PID;

   binnum = 1 + floor(timeData(:) / 39);
   Task1.Amp(f).trial = accumarray( binnum, ampData(:), [], @(V) {V})
   Task1.Time(f).trial = accumarray( binnum, timeData(:), [], @(V) {V})

   numtrial = numel(Task1.Amp(f).trial)
%    for trialnum = 1:numel(Task1.Amp(f).trial)
binnum(end)

for trialnum = 1:binnum(end)

       for numbers = 1:numel(Task1.Amp(f).trial{trialnum})

     timedata = Task1.Time(f).trial{trialnum}(numbers);
ampdata = Task1.Amp(f).trial{trialnum}(numbers);

       DATA_raw(rownum,1) = PID;
       DATA_raw(rownum,2) = trialnum;
       DATA_raw(rownum,3) = timedata;
       DATA_raw(rownum,4) = ampdata;

       rownum = rownum + numbers;
       end
end
end

The output data file DATA_Raw looks like this...

2  3  114.080000000000  0.218000000000000
0  0  0  0
0  0  0  0
0  0  0  0
2  4  120.260000000000  0.0350000000000000

Where are these zero's coming from???

Thanks, ML

[dpb]

Where is there an "output data file" or do you just mean the array in memory?

W/o any data to be able to see for certain it's not possible to say for absolute certain, but a reasonable conjecture is that your accumarray indices aren't what you think they are.

Best recommendation I can give is to use the debugger and step thru your code watching what happens as you process the data. I'll wager you'll find an "ah-ha!" moment when all becomes clear...now that may engender some other questions on how to best handle what you find, but I'll bet you find out the "why". :)

[Mary]

Thank you I'll give debug a try!

Answer 1

[Joseph Cheng]

I am suspecting that it is coming from your rownum = ronum+numbers.

From your snippit rownum starts at 1 and then gets incremented by numbers. It's basically defining a row in DATA_raw and then defining another row at (rownum+numbers) further down for the next iteration. since numbers isn't 1 you won't get sequential rows of data. When that happens MATLAB will zero fill the empty spaces.

simple display of whats happening

    x=rand(1,4)
    x(5,:)=rand(1,4)

I see where this may have worked if you get PID, trailnum, etc as size numbers*1 matrix where you'd need to increment the indexing by numbers and not just 1. However you'll additionally need to DATA_raw(rownum:rownum+numbers-1,1) which will get a numbers*1 matrix to go into that space.

[Mary]

Thank you! The example you gave was really helpful - I should have realized that sooner.

Hopefully fixing the way numbers is called will clean things up!

Thank you for such a quick and thoughtful response!