Calculate duration from labeled timestamped data.

Hi,

I have a timetable with timestamps and labels of different activities. The stacked plot looks like this.

I would like to calculate the durations of each cycle of the activity. For example, the top plot has 11 "Dump" sequences, so the outcome should be,

Dump = [t1, t2, ........t11], t = duration of each dump instance.

Is there any smarter way to accomplish this without using FOR loop and tracking changes in each row?

I am also attaching the .mat file of the timetable.

Thank you!

[Akira Agata]

If possible, could you upload your timetable data so that people here can try their idea?

[Khandakar Rashid]

Thank you Akira. I have uploaded the timetable data in the original question as "data.mat".

Answer 1

[Akira Agata]

Thank you for providing your data.

I believe the following is an possible solution. I hope this will be somehow helpful for you!

load('data.mat');

% Find start/end time for each 'Dump' period
idx = data.Truck1 == 'Dump';
locStart = find(diff(idx) == 1)+1;
locEnd = find(diff(idx) == -1);
timeStart = data.Time(locStart);
timeEnd = data.Time(locEnd);

% Calculate duration for each 'Dump' period
timeDuration = timeEnd - timeStart;

% Arrange to table
tblTruck1 = table(timeStart,timeEnd,timeDuration);

>> tblTruck1

tblTruck1 =

11×3 table

timeStart timeEnd timeDuration

_________ ________ ____________

06:44:46 06:45:44 00:00:58

06:56:53 06:57:48 00:00:55

07:11:13 07:12:21 00:01:08

07:27:06 07:28:15 00:01:09

07:38:22 07:39:19 00:00:57

07:52:06 07:53:02 00:00:56

08:01:43 08:02:30 00:00:47

08:12:52 08:13:39 00:00:47

08:28:23 08:29:19 00:00:56

08:40:42 08:41:35 00:00:53

08:54:25 08:55:23 00:00:58

[Khandakar Rashid]

Thank you so much Akira. Although I figured out a way to do this, your solution is much more elegant. I am going to use it :D :D