hourly average of multiple column data

I have made a code to calculate hourly average of minute data. But it works only for single column. I have multiple column of concentration data. How can i get the desired answers. pls help.

day1hour=hour(date==1);
day1bc=BC(find(hour(date==1)));
maxday1=max(day1hour);
for i=1:max(day1hour)
ll= find(day1hour==i);

    fffff=BC(ll);
    mean_BC_hourly1(i,:)=mean(fffff);
  end

The input datas are:

day  hour   Con 1  con2
1  1  10  20
1  1  20  40
1  2  30  60
1  2  40  80
1  3  50  100
1  3  60  120
1  3  70  140

The answer expected:

con1avg   con2avg
15  30
35  70
60  120

[Guillaume]

Note that none of the find in your code are necessary. They just slow down the processing.

day1bc = BC(hour(date == 1))); %works just as well

[Star Strider]

See also: http://www.mathworks.com/matlabcentral/answers/283934-hourly-average-of-multiple-column-data that seems to be an essential duplicate.

Answer 1

[Guillaume]

Use accumarray:

m = [1  1  10  20 
     1  1  20  40
     1  2  30  60
     1  2  40  80
     1  3  50  100
     1  3  60  120
     1  3  70  140]

[dayhour, ~, destrow] = unique(m(:, [1, 2]), 'rows');
subs = [repmat(destrow, 2, 1), repelem([1;2], numel(destrow))];
hourlymean = accumarray(subs, reshape(m(:, [3, 4]), [], 1), [], @mean);
array2table([dayhour, hourlymean], 'VariableNames', {'day', 'hour', 'con1avg', 'con2avg'})

Or the new split apply workflow introduced in R2015b

[navan]

Wow. Wonderful. Thanks a ton Guillaume. It is what i wanted really. It is a great help. thank you once again :) But if i have around 10 concentration how should i modify the code ?

[Guillaume]

colstoaverage = 3:12;
subs = [repmat(destrow, numel(colstoaverage), 1), repelem((1:numel(colstoaverage))', numel(destrow))];
hourlymean = accumarray(subs, reshape(m(:, colstoaverage), [], 1), [], @mean)