adding all fields of a structures

I have a following structure with fields:

say, my_struct = struct with fields:

bin_1: [851×19 table]

bin_2: [868×19 table]

bin_3: [164×19 table]

bin_4: [104×19 table]

bin_5: [348×19 table]

bin_6: [406×19 table]

bin_7: [303×19 table]

bin_8: [323×19 table]

bin_9: [275×19 table]

bin_10: [110×19 table]

I want to have a structure with one field that is all field combined together. Something like my_new_struct= bin_1: [3500(approx.)×19 table]

i.e. my_new_struct= sum of all tables of my_struct

Could someone help me with the program for that?

[Stephen23]

This is fragile data design: as Voss correctly pointed out, the order of fields may not be the order you expect or require. The cause is because you have hidden meta-data (i.e. pseudo indices) in the fieldnames. Much better data design would use actual indices (rather then pseudo-indices) e.g. in a simple structure array or a cell array. That also significantly simplifies the code required to merge all of the tables:

vertcat(C{:}) % from a cell array
vertcat(S.F) % from a structure array

Better data design leads to much better code.

Answer 1

[Matt J]

Tables=arrayfun(@(f)mystruct.(f)  ,"bin_"+(1:10) ,'uni',0);

my_new_struct.Bin=vertcat(Tables{:});

[Dyuman Joshi]

Using fieldnames would be better imo.

[Matt J]

If they are the only fields present, and in the order shown, then yes, fieldnames would probably be better.

Answer 2

[Voss]

Here's one way to do that:

% a structure like your "my_struct", but with only 3 fields:
my_struct = struct( ...
    'bin_1',array2table(repmat("bin_1_data",2,2)), ...
    'bin_2',array2table(repmat("bin_2_data",4,2)), ...
    'bin_3',array2table(repmat("bin_3_data",3,2)))

my_struct = struct with fields:
    bin_1: [2×2 table]
    bin_2: [4×2 table]
    bin_3: [3×2 table]

% combine all the tables:
C = structfun(@(x){x},my_struct);
T = vertcat(C{:})

T = 9×2 table
        Var1            Var2    
    ____________    ____________

    "bin_1_data"    "bin_1_data"
    "bin_1_data"    "bin_1_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_3_data"    "bin_3_data"
    "bin_3_data"    "bin_3_data"
    "bin_3_data"    "bin_3_data"

However, that may not combine the tables in the order you wanted. In order to impose an order, you can do something like this:

% get the names of the fields:
fn = fieldnames(my_struct);

% reorder them somehow:
% fn = sort(fn);   % sorted lexicographically
% fn = natsort(fn) % sorted by the number they contain
fn = fn(end:-1:1); % reverse

% combine all the tables in that order:
C = cellfun(@(f)my_struct.(f),fn,'UniformOutput',false);
T = vertcat(C{:})

T = 9×2 table
        Var1            Var2    
    ____________    ____________

    "bin_3_data"    "bin_3_data"
    "bin_3_data"    "bin_3_data"
    "bin_3_data"    "bin_3_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_2_data"    "bin_2_data"
    "bin_1_data"    "bin_1_data"
    "bin_1_data"    "bin_1_data"

See natsort in the File Exchange, which may give you the ordering you're after in this case.

https://www.mathworks.com/matlabcentral/fileexchange/34464-customizable-natural-order-sort

Answer 3

[Bruno Luong]

s.bin_1=array2table(rand(2,3));
s.bin_2=array2table(rand(3,3));
s.bin_3=array2table(rand(4,3));

c=struct2cell(s);
newstruct.T = cat(1,c{:});

disp(newstruct.T)

     Var1        Var2       Var3  
    _______    ________    _______

    0.90585     0.16903    0.21998
     0.6887     0.44019    0.43033
     0.3722     0.79743    0.50584
    0.64057     0.40379    0.56822
    0.14505      0.4095    0.27283
    0.71969    0.053031    0.60168
    0.67643      0.2598    0.83108
    0.17963     0.71627    0.46944
    0.77515     0.42908      0.429