How to extract cell array of 2D matrices into a timeseries?

I have some data that is stored as a 1666x1 cell array. Each cell contains a 4x19 double. I need to extract this data and attach it to a timeseries object. Timeseries will not take a cell array directly, so I need to extract the cells.

So let's say the cell array is named data, and the timeseries is ts. I need:

ts.Data(1) = data{1}

ts.Data(2) = data{2}

etc.

cell2mat() concatenates everything into 6664x19, and I can't get reshape() to chop it up into the correct 4x19 blocks. I could do it in a for loop, but that's not very satisfying in MATLAB.

I know this is a simple syntax thing, but I can't seem to find an answer that works.

[Burak]

To populate a timeseries object with data from a cell array where each cell contains a 4x19 double, you can indeed use a for loop. However, if you want a more MATLAB-esque solution (avoiding explicit loops), you can use arrayfun to achieve this. Here’s how you can do it:

   % Initialize the timeseries object with an empty array of the correct size
    ts = timeseries();
    % Use arrayfun to attach data to the timeseries object
    ts.Data = cat(3, data{:});

Answer 1

[Walter Roberson]

data = {rand(2,5); rand(2,5); rand(2,5)};
ts = timeseries(cat(3,data{:}))

  timeseries

  Common Properties:
            Name: 'unnamed'
            Time: [3x1 double]
        TimeInfo: tsdata.timemetadata
            Data: [2x5x3 double]
        DataInfo: tsdata.datametadata

getsamples(ts,1)

  timeseries

  Common Properties:
            Name: 'unnamed'
            Time: [1x1 double]
        TimeInfo: tsdata.timemetadata
            Data: [2x5 double]
        DataInfo: tsdata.datametadata

So in general you

ts = timeseries(cat(3, data{:}));

[Daniel Flisek]

That worked! I knew it'd be something simple. Thanks for taking the time to answer.