列的总和取决于条件语句Matlab

Question

I have a matrix in matlab where I want to sum the values of a column if other columns check with a conditional statement, then I'd like to store the summed values in a coherent way.

lay-out input table ListMax:

year    Month    day    hour    precipitation  
1998    1        1      1       5
1998    1        2      2       7
1998    1        3      3       0
....    ...     ...    ...      ...

lay-out output result_matrix:

year    jan    feb   mar    
1998    100    120   140
1999    90     110   130
...    ...     ...   ...

I'm struggeling with the combination of itirating over the rows, checking the conditional statement and saving the values in an output.

I had the following in mind, but it's not working:

result_matrix = zeros(15,12);   %empty matrix 15 years, 12 months
year_number = 1998;
counting_years = 1;
counting_months = 1; 

for (ii = 1: length(ListMax));
    if ListMax(:,1) == year_number && ListMax(:,2) == counting_months;
        result_matrix(counting_years, counting_months) = (sum(Listmax(:,5));
    else % update month itirators
        if counting_months < 12
            counting_months = counting_months + 1;
        else % end of year, set month count to 1
            counting_months = 1;
        end
     year_number = year_number +1;
     counting_years = counting_years + 1;
     end
end

I can see this is probably not the most straightforward approach, but at the moment this the approach I can think of, now I just need to get it working yhough. Any push in the right direction would be much appreciated :)

Answer 1

You can simply use accumarray with a two columns sub index.

year    Month    day    hour    precipitation  
1998    1        1      1       5
1998    1        2      2       7
1998    1        3      3       0
....    ...     ...    ...      ...

For example, if you want to sum the total amount of precipitation according to the year and to the month, you can use the two first column of your matrix as index and the last one as precipitation's value.

sub = 1998    1     %Jan.1998  
      1998    1     %Jan.1998   
      1998    2     %Feb.1998 
      ...
sub(:,1) = sub(:,1) - (min(sub(:,1))-1); %so the index for the year start at 1 and not 1998.

val = 6 %millimeter of precipitation on Jan.1998 day xxx
      4 %millimeter of precipitation on Jan.1998 day xxx
      7 %millimeter of precipitation on Feb.1998 day xxx
      ....

And now use accumarray :

result = accumarray(sub,val,[],@sum);

Answer 2

How about this (untested):

year_start = 1998;
year_end = 2016;
result_matrix = zeros(year_end-year_start+1,12);

for year = year_start:year_end
    for month = 1:12
        rows = (ListMax(:,1)==year) & (ListMax(:,2)==month);
        result_matrix(year-year_start+1,month) = sum(ListMax(find(rows),5));
    end
end