Help in averaging climate data

I am a novice in matlab and looking for help with calculating average monthly data from hourly data.

I have the data for full year in one file and have multiple txt files for different years. Column 1 is year, column 2 is month of year (1 to 12), column 3 is day of month and column 4 is hour of day (0 to 23). Looking forward to some guidance on how to do this.

Thanks in advance

[Kye Taylor]

Have you imported the data into MATLAB? I.e. do you have the data stored in a N-by-4 dimensional matrix yet?

[Kye Taylor]

So, does a row of the data looks like

[1990, 2, 28, 23] % for 11pm Feb 28 1990

Where is the climate data? Is that stored in another column? or a separate vector?

[Sam]

This is how I imported the data to matlab:

D = importdata('metdata1.txt')

D =

          data: [8760x13 double]
      textdata: {8760x1 cell}
    rowheaders: {8760x1 cell}

[Kye Taylor]

So, which columns of the 13 are the four you describe above, and which one is the climate data?

[Sam]

And this is what the first row looks like if I am doing it right:

 D.data(1,:)

ans =

   1.0e+03 *

Columns 1 through 12

    1.9730    0.0010    0.0010         0   -0.9990   -0.0999   -0.0090   -0.9999   -0.9999   -0.9999   -0.9999   -0.0990

Column 13

   -0.0990

Answer 1

[Sam]

Yes, I have imported them into matlab

Answer 2

[Roger Stafford]

I'll suppose you have the climate data in column five of an array, A, with the year, month, day, and hour in its columns 1, 2, 3, and 4, respectively.

 [u,~,ix] = unique(A(:,1:2),'rows');
 B = [u,accumarray(ix,A(:,5))./accumarray(ix,1)];

[Sam]

Hi Roger, Thanks, this seems to produce a matrix for the 12 months. But how not to take into account the "999.9" data i.e. the missing value data in the column?

[Roger Stafford]

It should produce a row for every year-month combination, not just for 12 months.

If you can fully(!) describe the "missing value" situation, perhaps a solution could be found. As things stand, I can only guess at what you mean.

[Sam]

Actually data for different years are in different files in the data column, a data value of "999.9" means a missing value so not to be considered in the a calculation. Now it is averaging over the entire column for unique values of month.

[Roger Stafford]

You could initially remove all rows with the "999.9" data values:

 A = A(A(:,5)~=999.9,:);

However, it might be better to allow some tolerance for rounding differences:

 A = A(abs(A(:,5)-999.9)<100*eps(999.9),:);

[Noor Mohd Safar]

Thank you guys.