How to search a field in a structure and extract all fields that match

Hopefully I'll explain clearly as I haven't worked with structures much. I have been given a large 1x1 structure of arrays with numerous fields. I want to search one field to match a specific string. I then want to extract from the structure all fields for those cases where the string matched. I would prefer to do this with out loops. For an array of structures, I've been able to use logical indexing to access and extract data using "index2 = cellfun(@(x) any(strcmp(x, test)), {diab2.preg})" and then just index into the structure, but this doesn't work for a structure of arrays.

Answer 1

[Sindhu Priya]

Hi Dave,

I think, the following code snippet will help you.

s.f1= {'Sunday'  'Monday'}

s.f2= {'Sunday'  'Monday'  'Tuesday'}

s.f3= 'Tuesday'

s.f4= 'Wednesday'

s.f5= 'Thursday'

s.f6: 'Friday'

s.f7: 'Saturday'

index2 = structfun(@(x) any(strcmp(x, 'Monday')),s)

and the generated Output is

index2 =

    7×1 logical array

     1
     1
     0
     0
     0
     0
     0

Please refer this link for more details about 'structfun'.

Regards,

Sindhu

[Dave]

Thanks Sindhu, from this and continuing to research structures I realize the data I'm given is a structure of arrays. As an example the structure data is made up of

data.Sen = {'U1'; 'U2'; 'U1'; 'U1'}
data.Tid = {'1'; '1'; '1';'2'}
data.Obj = {'U1_1';'U2_1';'U1_2';U1_1'}
data.X   = {'10'; '5'; '3'; '1'}
data.Y   = {'20'; '7'; '4'; '2'}
I've sort of brute forced a way to get the related data I want with

     senidx = strcmp(data.Sen,'U1');
     thesen = data.Sen(senidx)
     theX   = data.X(senidx)
     theY   = data.Y(senidx)
     theTid = data.TID(senidx)

 but I expect there is a more elegant way to do this.

I thought the following would work
for nField = 1:numel(myfieldnames)
  mField = myfieldnames{nField}
  mydata.(mField) = data.(mField)(strcmp(data.Sen_T,'U1'))
end

but I get this error: Struct contents reference from a non-struct array object.

I think I'm close so any further insight would be appreciated.

Answer 2

[George Abrahams]

Old question, but for future reference, most likely you wanted to use a table, not a structure of cell arrays, as this makes it easy.

data.Sen = { 'U1'; 'U2'; 'U1'; 'U1' };
data.Tid = { '1'; '1'; '1'; '2' };
data.Obj = { 'U1_1'; 'U2_1'; 'U1_2'; 'U1_1' };
data.X   = { '10'; '5'; '3'; '1' };
data.Y   = { '20'; '7'; '4'; '2' };

% Convert structure of cell arrays to table (fields become columns).
data = struct2table( data );
% Find logical indices of the matching table rows.
% If you don't care about case or additional letters, see: strncmp,
% strcmpi, strncmpi.
rows = strcmp( data.Sen, 'U1' );
% Extract these rows from the table.
data( rows, : )

% ans = 3×5 table
%      Sen       Tid       Obj         X         Y   
%     {'U1'}    {'1'}    {'U1_1'}    {'10'}    {'20'}
%     {'U1'}    {'1'}    {'U1_2'}    {'3' }    {'4' }
%     {'U1'}    {'2'}    {'U1_1'}    {'1' }    {'2' }

If you didn't want to do that for some reason, you could use my fieldfun function on File Exchange / GitHub, which process the fields of one or more structures and outputs a structure with the same fields.

idx = strcmp( data.Sen, 'U1' );
fieldfun( @(x) x(idx)', data )

% ans = struct with fields:
%     Sen: {'U1'  'U1'  'U1'}
%     Tid: {'1'  '1'  '2'}
%     Obj: {'U1_1'  'U1_2'  'U1_1'}
%       X: {'10'  '3'  '1'}
%       Y: {'20'  '4'  '2'}