Is there a way to optimize this?

I'd like to speed up this code but I don't know how to improve it more, is there a way to optimize it?

I need to build an array with every Status On for every minute of the day.

Date = datetime(2018,11,08);
Time_Ref = Date+minutes(1):minutes(1):Date+days(1);
tres = cell(length(Time_Ref),1);
Time_on = {'2018-11-08 00:00:00';'2018-11-08 00:00:00';'2018-11-08 00:00:00';'2018-11-08 00:46:29';...
    '2018-11-08 03:51:33';'2018-11-08 08:23:40';'2018-11-08 23:51:37'};
Time_off = {'2018-11-09 00:00:00';'2018-11-09 00:00:00';'2018-11-09 00:00:00';'2018-11-08 12:46:41';...
    '2018-11-08 03:52:44';'2018-11-08 08:23:50';'2018-11-08 23:51:47'};
Status = {850;3030;6042;6052;7018;6052;6052};
tic
for j=1:length(Time_Ref)
    n=1;
    while true
        if Time_off(n)> Time_Ref(j)-seconds(59) && Time_on(n)<Time_Ref(j)
            tres{j} = [tres{j},Status(n)];
        end
        if n >= length(Time_on) || Time_on(n) > Time_Ref(j)
            break;
        end
        n=n+1;
    end
    
end

Answer 1

[Deepak]

Hi @Jonathan Pelletier-Marcotte, from my understanding, you have generated the “tres” array that stores the status codes for each minute of the day, based on when each status is active. You have built the “Time_Ref,” “Time_on,” and “Time_off” arrays, which serve as input to the for loop. Now, you want to optimize the code further to get the same result.

To do this, we can convert “Time_on” and “Time_off” to datetime objects once to avoid repeated conversions inside the loop. Next, we can pre-allocate the “tres” array with empty arrays, which improves the performance. Finally, we can use logical indexing inside the for loop to avoid nested loops that improves the performance further.

Below is the updated MATLAB code with the described changes:

Date = datetime(2018, 11, 08); 
Time_Ref = Date + minutes(1):minutes(1):Date + days(1); 
tres = cell(length(Time_Ref), 1); 

% Convert Time_on and Time_off to datetime arrays 
Time_on = datetime({'2018-11-08 00:00:00'; '2018-11-08 00:00:00'; '2018-11-08 00:00:00'; '2018-11-08 00:46:29'; ... 
    '2018-11-08 03:51:33'; '2018-11-08 08:23:40'; '2018-11-08 23:51:37'}); 

Time_off = datetime({'2018-11-09 00:00:00'; '2018-11-09 00:00:00'; '2018-11-09 00:00:00'; '2018-11-08 12:46:41'; ... 
    '2018-11-08 03:52:44'; '2018-11-08 08:23:50'; '2018-11-08 23:51:47'}); 

Status = [850; 3030; 6042; 6052; 7018; 6052; 6052]; 

tic 
for j = 1:length(Time_Ref) 
    % Find indices where the condition is met 
    indices = Time_off > Time_Ref(j) - seconds(59) & Time_on < Time_Ref(j); 

    % Collect statuses for those indices 
    tres{j} = Status(indices)'; 
end 
toc 

Elapsed time is 0.082180 seconds.

Attaching the documentation for logical indexing in MATLAB for reference:

https://www.mathworks.com/help/matlab/matlab_prog/find-array-elements-that-meet-a-condition.html

I hope this helps.