Check if datetime is within a time interval and add to sequence accordingly?

EDIT: I'm now looking at the isbetween function in MATLAB

___________________

Hi all,

I'm a bit stuck on trying to build a sequence of events based on a truncated interval.

I have a vector, t3, containing datetimes and a vector, flags, containing sequence identifiers (1 for system up and 2 for down). I also have a vector, tInts, of datetimes for 5 minute intervals that span the entirety of t3.

So, for example, I have:

'2019-01-24 16:00:00'
'2019-01-24 16:00:04'
'2019-01-24 16:02:02'
'2019-01-24 16:03:36'
'2019-01-24 16:05:44'
'2019-01-24 16:06:12'
'2019-01-24 16:08:12'
'2019-01-24 16:09:13'
'2019-01-24 16:14:04'
'2019-01-24 16:14:35'

for t3,

flags = [2;1;2;1;2;1;2;1;2;1];

and tInts would be

'2019-01-24 16:00:00'
'2019-01-24 16:05:00'
'2019-01-24 16:10:00'
'2019-01-24 16:15:00'

I want to check that the values of t3 are within the 5 minute intervals in tInts, and, if they roll over, I want to insert an additional flag into a sequence. So, based on my example, at index 5 for t3, we exceed 5 minutes. The flags to that point are [2;1;2;1;2]; Now, I want to add an additional 1 (since it rolled over the 5 minutes in state 1) before the last 2, making

seq = [2;1;2;1;1;2]

and on and on through all of t3.

So I have a thought to try an if-statement inside a for-loop where I have some condition that would look at a certain t3 element and check if the difference from tInts(j) - t3(1) >= 0 then seq = [flags upto i; flags(i-1)] else seq = flags up to i.

I'm not really sure. Any help would be great.

Thanks,

E

Answer 1

[Erin Winkler]

Hi all,

I actually figured it out on my own. For anyone who needs it in the future (I was using it for building a sequence for a Transition Matrix for a Markoc Chain), here's how I did it:

seq = []; 

for i = 1:length(tInts)-1
        
        tf = isbetween(t3,tInts(i),tInts(i+1));
        indxF = find(tf,1,'first');
        indxL = find(tf,1,'last');
        
        seq = [seq; flags(indxF:indxL); flags(indxL)];
        
 end
    

Answer 2

[Abhishek Kumar Tiwari]

Hi guys, I am working on my thesis and need some help.

In my project,

x=1:number_of_signals

Signal(x).timestamp:

• contains the timestamp at which the signal x acquisition took place (datetime type); or, the frequency at which signal x was acquired by the sensor
• It may be have flexible number of elements i.e. (50000 x 1) or (100 x 1) data points

'30-Mar-2018 14:18:48'
'30-Mar-2018 16:32:28'
'30-Mar-2018 16:56:16'
'30-Mar-2018 16:57:30'
'30-Mar-2018 16:58:28'
'30-Mar-2018 16:59:25'
'30-Mar-2018 16:59:54'
'30-Mar-2018 17:00:51'
'30-Mar-2018 17:01:49'
'30-Mar-2018 17:02:46'

...............etc......

Signal(x).values:

• contains values corresponding to timestamps mentioned above (double type)

timestamp:

• contains timestamp (datetime type)
• It has fixed number of elements i.e. (2000 x 1) data points

'30-Mar-2018 16:40:08'
'30-Mar-2018 16:40:09'
'30-Mar-2018 16:44:37'
'30-Mar-2018 16:44:37'
'30-Mar-2018 16:44:38'
'30-Mar-2018 16:56:15'
'30-Mar-2018 16:57:00'
'30-Mar-2018 16:57:29'
'30-Mar-2018 16:57:58'
'30-Mar-2018 16:58:27'
'30-Mar-2018 16:58:56'

...............etc......

How can I fit the values of Signal(x).timestamps in the time intervals of timestamps which are fixed?

If there are more than 1 values for the time interval, we can take the average of values for each time interval.

If there are no values for the time intervals, we can put 0 corresponding to that time interval.

Please help.