Timeseries syncronized by time

Is it possible to match 2 or more timeseries that have different time vector, syncronized by time?

For example,let's say, ts1 goes form 07:00:00 to 09:00:00 o'clock, ts2 goes from 08:00:00 to 09:00:00, ts3 from 08:30:00 to 10:00:00. each data should be visible just in the right time they were taken:in this way I could analyze data for a single interval of time from different instrumentations...

or maybe it is better to create a single timeseries with 3 different dataset?

thanks!

Answer 1

[dpb]

Convert all to datenums with datenum and then you can

a) plot on a common axis and automagically they'll line up correctly (use datetick to label the axes in time), and

b) use logical comparisons to select subsets of a common start/end time within the overall time vectors.

"Piece o' cake"... :)

[Engineer]

Thanks! I was trying to use "timeseries" function in order to avoid playing with strings and dates! The problem is the different number of dimensione (rows) of the two datatset (because the first one is longer starting before)

St5=PT(:,8:end);

St5_ts = timeseries(St5,1:numel(St5(:,1)),'Name','PT5');

St5_ts.TimeInfo.Increment = 1/4;

St5_ts.TimeInfo.Units = 'seconds';

St5_ts.TimeInfo.Format = 'dd-mmm-yyyy HH:MM:SS';

St5_ts.TimeInfo.StartDate = '12-Dez-2013 07:29:59';

PT=txt2mat('St6.txt');

St6=PT(:,8:end);

St6_ts = timeseries(St6,1:numel(St6(:,1)),'Name','PT6');

St6_ts.TimeInfo.Increment = 1/4;

St6_ts.TimeInfo.Units = 'seconds';

St6_ts.TimeInfo.Format = 'dd-mmm-yyyy HH:MM:SS';

St6_ts.TimeInfo.StartDate = '12-Dez-2013 09:29:59';

%%now, I'd like to "connect" all the timeseries in a only one, or having a single timeseries with all data in it. Thus,getting data from entire timeseries containing all data

dd=getsampleusingdata(timeserieswithallinstruments,'chosentime')

X=dd.data % from where I could have all the data for the chosen interal

Do you think would be better working with normal arrays?

[dpb]

Quite possibly (probably, even, maybe?) yes.

I've not played with the timeseries object enough to be very familiar with it, but what little I have it seems the only way to do what you're asking is to essentially build the time vector into it manually as it just uses ordinal number by default (or a fixed origin, fixed dt). Perhaps you could figure out a way to do that by picking the first origin and the last and blank filling to make them all commensurate in length but that seems excessive storage for nothing to me.

I'm not certain what your request above means of "having a single time seres with all data in it". If the times are duplicated or the data is from different instruments, do you really want to commingle it into one long vector? Not knowing what you then intend to do with it, I don't know how to better answer, specifically.

[Engineer]

ok, thanks for your time!