Find cell position of a timeseries array

Hello all,

we need to find the cell position of a timeseries which contains the time-value we are looking for. There is a big 1x1 double timeseries called test.

test.Time has all time-values and test.Data got signal-values. We need a start- and end-time for example: second 5 and second 10.

How is it possible to find the location of those cells which contains values close to 5 and 10. The timeseries are very big that is why maybe the time-values 5 or 10 are not in there but 5.025, 5.0042 or 10.05796 are possible to find.

A solution without a for loop is prefered.

Thank you for any help!

Answer 1

[Meg Noah]

Assuming your data are in vector arrays that are, say, doubles, and packed as not cell arrays:

test.Time = 0:0.1:50;
test.Time = test.Time + 0.001*rand(size(test.Time));
test.Data = sin(2*pi*test.Time/25) + 0.3*rand(size(test.Time))-0.3*0.5;
startTime = 5;
endTime = 10;
idxStart = find(test.Time >= startTime,1,'first');
endTime = find(test.Time > endTime,1,'first') - 1;
figure()
hold on; ylabel('Data'); xlabel('Time (s)'); box on;
plot(test.Time,test.Data,'DisplayName','Data');
plot(test.Time(idxStart:endTime),test.Data(idxStart:endTime), ...
    'DisplayName','Time Wanted');
legend('location','best');
plot([test.Time(idxStart) test.Time(idxStart)],[-1.5 1.5],':', ...
    'DisplayName',['Start time = ' num2str(test.Time(idxStart)) ' s']);
plot([test.Time(endTime) test.Time(endTime)],[-1.5 1.5],':', ...
    'DisplayName',['End time = '  num2str(test.Time(endTime)) ' s']);

Note: Indexing without using find is faster and better, but since you wanted to see the actual times of the start and stop, I used a find command.

[Ali Sahouli]

Thank you very much!

[Meg Noah]

You're very welcome. I should have mentioned that it's important to organize the time data monotonically increasing for this to work. You can use the sort function something like:

% test if the data are monotonically increasing
function tf = mono_increase(x)
  tf = all(diff(x)>0);
end

% if not, sort the data
% find the index order that will result in monotonically increasing data
[~,idx] = sort(test.Time);
% then reorganize the data
test.Time = test.Time(idx);
test.Data = test.Data(idx)