Creating table with array in element

I am trying to create a table which includes an array in each row (like the output from regionprops etc.)

I know how to do it by creating a struct and using struct2table, but is there are way to create a blank table with an array variable using the table('Size', [..], 'VariableTypes', ...) constructor?

I have tried searching but any useful results are buried under posts about array2table.

Answer 1

[Walter Roberson]

You do not need to do anything special. Once the variable exists, you can write over the content with an array

t = table('Size', [3 2], 'VariableTypes', {'uint8', 'double'})

t = 3×2 table
    Var1    Var2
    ____    ____

     0       0  
     0       0  
     0       0  

t.Var2 = zeros(3, 2)

t = 3×2 table
    Var1     Var2 
    ____    ______

     0      0    0
     0      0    0
     0      0    0

The only challenge would be to have a variable created in the table to have an initial size that has multiple columns; there is no provision for doing that as one step (other than calling table() and passing in content of the appropriate size.)

[Philip Shaw]

I see, I was trying to assign the individual elements and so getting a size mismatch.

I suppose that means there's no way to put a multi-row matrix in a table element, apart from wrapping it in a cell or reshaping it so the first dimension has length 1.

[Walter Roberson]

You can see that regionpops does it by using cell() for the multi-row entries.

img = imread('cameraman.tif');
bw = img > 224;
imshow(bw)

props = regionprops('table', bw, 'area', 'centroid', 'extrema')

props = 31×3 table
    Area        Centroid          Extrema   
    ____    ________________    ____________

     28     106.61    173.61    {8×2 double}
      1        103        76    {8×2 double}
     69     105.81    242.49    {8×2 double}
      5        108       230    {8×2 double}
      1        111       215    {8×2 double}
      6     112.83    206.33    {8×2 double}
      3        112       210    {8×2 double}
      1        116       190    {8×2 double}
      1        119       175    {8×2 double}
      3        130        89    {8×2 double}
     89     140.51    95.562    {8×2 double}
     46      135.2    143.52    {8×2 double}
      2      135.5       165    {8×2 double}
      1        135       180    {8×2 double}
     18        135     190.5    {8×2 double}
      2        137     114.5    {8×2 double}

varfun(@class, props)

ans = 1×3 table
    class_Area    class_Centroid    class_Extrema
    __________    ______________    _____________

      double          double            cell