Average data in multiple excel file using MATLAB

Question 1

The following code might serve your requirements. Again, since you are interested in daily averages, the output would be much be smaller, as it's over a 24 hour period and I am assuming you need that. Also, it takes care of your missing data condition.

Code -

%% Setup params and data
start_hour = 17;
[num,txt,raw] = xlsread('summary_file.xls');

datenums = NaN(size(num,1),1);
for count = 1:size(num,1)
    year1 = cell2mat(raw(count,1));
    date1 = cell2mat(raw(count,2));
    time1 = cell2mat(raw(count,3));
    date_str = strcat( num2str(year1) , '-', date1 );
    datenums(count) = datenum( date_str, 'yyyy-dd-mmm')  + time1(:);
end

%% Take care of conditions
firstdata_start_hour = round(24*cell2mat(raw(1,3)));
if firstdata_start_hour > 17
    start1 =  floor(datenums(1))  + (start_hour/24);
elseif firstdata_start_hour < 17
    start1 =  floor(datenums(1))-1  + (start_hour/24);
else
    start1 = datenums(1);
end

ind1 = floor(datenums-start1) + 1;

%% Start Processing
num_items = size(num,2)-3;
num_days = max(ind1);

bins = NaN(num_days,num_items);
for count1 = 1:size(bins,2)
    for count2 = 1:size(bins,1)
        bins(count2,count1) = mean(num(find(ind1==count2),count1+3));
    end
end
bins(isnan(bins))=0;
average_nums = bins

Output of the averages for some data, compiled by OP on my request -

+-----------+---------------+------+-------------+------------+-----------+--------------+
|   Date    | Pressure(kPa) | Temp | Humidity(%) | W-spd(m/s) | Radiation | Rainfall(mm) |
+-----------+---------------+------+-------------+------------+-----------+--------------+
| 8/10/2009 | 100.1         | 25.8 | 79.1        | 1.4        | 82.6      | 1.7          |
| 8/11/2009 | 100.2         | 27.5 | 75.7        | 1.9        | 173.8     | 0.0          |
| 8/12/2009 | 100.1         | 28.4 | 73.5        | 2.1        | 177.1     | 0.0          |
| 8/13/2009 | 100.0         | 28.4 | 73.2        | 2.5        | 197.4     | 0.0          |
| 8/14/2009 | 100.0         | 28.5 | 73.5        | 2.2        | 151.2     | 0.0          |
| 8/15/2009 | 100.2         | 27.3 | 75.4        | 1.2        | 96.2      | 0.4          |
| 8/16/2009 | 100.2         | 27.1 | 75.5        | 1.4        | 122.6     | 0.0          |
| 8/17/2009 | 100.2         | 27.2 | 75.7        | 1.5        | 158.3     | 0.2          |
| 8/18/2009 | 100.2         | 27.5 | 72.2        | 1.4        | 186.4     | 0.0          |
| 8/19/2009 | 100.3         | 28.4 | 68.4        | 1.9        | 186.9     | 0.0          |
| 8/20/2009 | 100.3         | 28.1 | 69.2        | 2.0        | 184.8     | 0.0          |
| 8/21/2009 | 100.3         | 26.5 | 75.8        | 1.3        | 122.3     | 0.6          |
+-----------+---------------+------+-------------+------------+-----------+--------------+

For multiple excel files, you have to loop over all such files in a loop. There doesn't appear to have any other way out.

Question 2

When you use Matlab's xlsread it opens and closes the Excel COM Server, so using this with multiple files will slow down execution. Read this post at Matlab Central which explains how can you avoid opening and closing Excel COM Server.

As for the missing data, maybe you could interpolate 17:00 with the data from find(time<17:00,'last') and find(time>17:00,'first')

Question 3

I love using MATLAB but if you don't mind me asking, have you considered using the R statistical software Package? Its really cool in my opinion you don't have to agree with me.

It is very easy to do the tasks that you are asking for. R can even handle missing data and easily work around it and can read excel files easily and search for the data you want.

Where to get R:

http://cran.r-project.org/