MATLAB: Faster way to resample a timeseries

optimizationresampletimeseries

Hello,
I have about one month worth of data in a 1 minute timescale and I have a series of events which I need to resample onto that timescale.
 for m=1:length(UniqueEvents)
        EventLogical=strcmp(Log.Event, UniqueEvents{m});
        %create a time series object for each event
        EventSeries=timeseries(EventLogical, datestr(LogBook.EventStartTime));
        TimeSeries=resample(EventSeries, datestr(DateTable.TimeStamp), 'zoh');
        
        if m==1
            %convert time series into cell array
            TableSize=size(getabstime(TimeSeries));
            StatusTable=table(zeros(TableSize(1),1));
            StatusTable.TimeStamp=datetime(getabstime(TimeSeries));
            VarName=strcat('Status_',Var{k});
            StatusTable.Var1=[];
        end
        StatusTable.(VarName)(logical(TimeSeries.Data))=UniqueEvents(m);
 end
Using the profiler the vast majority of runtime is a result of the resampling line
         TimeSeries=resample(EventSeries, datestr(DateTable.TimeStamp), 'zoh');
Does anyone have any ideas on how to make this faster?

Best Answer

If you're using a release where timetable is available, I recommend storing your data in a timetable and using the retime function.
Related Question