How to add data stored in structure depending on other data related but stored in a different structure

Hi people!

I have some streamflow data that are stored in a structure called 'debit' (1x78888). Each streamflow data are related to a date and time (dd-MM-yyyy HH:mm:ss), which are stored in an other structure called 'Date' (1x78888). I want to have a mean streamflow per day. So what I did, is that I copied all the dates but without the hours to have data like 'dd-MM-yyyy'. Then I tried to do a while loop to sum all streams for a same day.

So far, my script looks like bellow, but it doesn't works. All I have in my structure called 'debit_total' are zeros...

Can someone help me with this ?

I hope it is clear enough !

load DE5.mat;
debit = Data.raw.horaire.debits;
date = Data.raw.horaire.time;
date.Format = 'dd-MMM-yyyy';
date_jour = date;
debit_total = zeros(size(debit));
i = 1;
while i <= length(date_jour)
    if date_jour(i) == date_jour(i+1)
        debit_total(i) = debit(i) + debit(i+1);
        debit_total(i+1) = debit(i) + debit(i+1);
    end
    i = i + 1;
end

[Stephen23]

Use MATLAB functions, rather than renventing everything yourself. Store the data in a table, then use one of these:

https://www.mathworks.com/help/matlab/ref/double.groupsummary.html

https://www.mathworks.com/help/matlab/matlab_prog/split-data-into-groups-and-calculate-statistics.html

Answer 1

[chicken vector]

I am not sure I udnerstand correctly what you need, because this line is confusing me:

debit_total = zeros(size(debit));

You want the average per day but you initialise the output with the same number of elements of every time instant.

Is something like this close to what you are looking for?

% Create fictitious data:
N = 5e3;
totDays = 40;
debit = randi(1e2, 1, N);
date = linspace(datetime('now'), ...
               (datetime('now') + days(totDays)), ...
               N) - years(1);

% Find of unique dates:
[y, m, d] = ymd(date);
[~, idx] = unique([y', m', d'], 'Rows', 'Stable');
uniqueDays = date(idx);

% Loop over each unique day:
meanDebit = zeros(size(uniqueDays));
for j = 1 : length(uniqueDays)
    
    % Find all debits corresponding to the same [day, month, year]_
    meanDebit(j) = mean(debit(     day(date) ==   day(uniqueDays(j)) ...
                               & month(date) == month(uniqueDays(j)) ...
                               &  year(date) ==  year(uniqueDays(j)) ));
    
    % If timespan is less than a year you can compare only days and months:
    % meanDebit(j) = mean(debit(     day(date) ==   day(uniqueDays(j)) ...
    %                            & month(date) == month(uniqueDays(j)) );
    
    % If timespan is less than a month you can compre only days:
    % meanDebit(j) = mean(debit( day(date) == day(uniqueDays(j)) ));
end

% Display result:
uniqueDays.Format = 'dd-MMM-yyyy';
table(uniqueDays', meanDebit', 'VariableNames', {'Date', 'Average Debit'})

output = 41×2 table
       Date        Average Debit
    ___________    _____________

    02-Jun-2022       49.459    
    03-Jun-2022       47.472    
    04-Jun-2022       48.456    
    05-Jun-2022       52.528    
    06-Jun-2022       51.272    
    07-Jun-2022       56.096    
    08-Jun-2022       53.712    
    09-Jun-2022       51.016    
    10-Jun-2022       48.336    
    11-Jun-2022       51.976    
    12-Jun-2022       51.312    
    13-Jun-2022       54.032    
    14-Jun-2022       50.472    
    15-Jun-2022       52.984    
    16-Jun-2022         53.8    
    17-Jun-2022        51.12    

[Cybelus]

Yea i don't know, I am new to MatLab and tried to get help with ChatGPT... But you understood well, thank you very much !!