Convert two cell arrays (time values and data values) to one cell array of timeseries

Hello all.

I have two cell arrays: t and y. Cell t contains the time values of 200 timeseries. Y contains the data values of 200 timeseries. The size of Y and t are 1x200. Each elemnt is a 1D vector.

How can I create a new cell array AA with a size of 1x200, whose elements are the reconstructed timeseries from elemnts of t and Y. Please note that "timeseries" are the matlab format of timeseries(datavals,timevals)

Answer 1

[Voss]

AA = cellfun(@(dd,tt)timeseries(dd,tt),Y,t,'UniformOutput',false);

[yamid]

Thanks. Could you please also let me know, how can I find the timeseries that only have one datapoint (i.e., only one timevalue and one datavalue) and how to remove them from the AA cell array?

[Voss]

You're welcome!

idx = cellfun(@(x)isscalar(x.Time),AA);
AA(idx) = [];

[yamid]

That's perfect, Voss. Thanks a lot. Just a last question:

After this opertion I am gonna resample all the timeseries in AA, so that all timeseries have the same time values [1:170]. I do this:

BB = cellfun(@(tsin)resample(tsin,1:170),AA,'UniformOutput',false);

How can I now create a double matrix CC whose columns are the data values of each timeseries in BB? (matrix CC has 170 rows and the number of columns depends on the number of timeseries)

[Voss]

CC = cellfun(@(tsin)tsin.Data(:),BB,'UniformOutput',false);
CC = [CC{:}];

[yamid]

Thank you, Voss. After resampling, the cell array BB contains several timeseries that have lots of NaN values (due to extrapolation). Is there a way to remove all those data points from each timeseries in BB that have NaN in Time and Data?

[Voss]

BB = cellfun(@(tsin)delsample(tsin,'Index',find(isnan(tsin.Data))),BB,'UniformOutput',false);