Create a struct inside a cell array

I'm using struct2xml from FileExhange and it's quite picky on the format of the structure to export. Basically I loose information when exporting data arrays (they are converted to strings). Example;
x = 1:3;
s.child.Text = x;
xml = struct2xml(s)
xml =
<?xml version="1.0" encoding="utf-8"?>
1 2 3
As you can see, the array of doubles has been converted into a string which makes it cumbersome for me to read the data back. I would like to have a format of;
<?xml version="1.0" encoding="utf-8"?>
1
2
3
but here's my pickle. In order for that to happen, I have to have a cell array of child-objects with a struct('Text', Value) in each.
s.child.Text = num2cell(x);
won't work, as struct2xml ultimately expects a struct (with field Text) but only gets a cell.
s.child = struct('Text', num2cell(x));
won't work, as that generates three separate structs. Multiple structs must be in cells.
cellfun(@struct, 'Text', num2cell(x));
won't accept the field name 'Text' unfortunately.
Any brilliant ideas?

 Respuesta aceptada

Jan
Jan el 4 de Jul. de 2012
Editada: Jan el 4 de Jul. de 2012
I do not know struct2xml, but the author does. Did you try to contact him already?
Intuitively I'd try:
s.child(1).Text = 1;
s.child(2).Text = 2;
s.child(3).Text = 3;

4 comentarios

No I did not. I see this as a general problem of getting the data to the desired format.
At the moment, I solved it by
for i=1:length(x)
s.child{i} = struct('Text', x(i));
end
But it's not very pretty.
Jan
Jan el 4 de Jul. de 2012
s.child = cell(1, length(x)); % Pre-allocate
for i = 1:length(x)
s.child{i}.Text = x(i);
end
Is this "nicer"?
Sebastian Holmqvist
Sebastian Holmqvist el 4 de Jul. de 2012
Your 3rd line won't give me the correct structure, but I appreciate the pre-allocation part. I guess I'll accept it for what it is. I was looking for some form of "vectorization" but it seems cumbersome.
Thanks!
Jan
Jan el 4 de Jul. de 2012
Editada: Jan el 4 de Jul. de 2012
Are you sure? It looks like my code creates exactly the same s as yours. Anyway, vectorization is efficient, when the data can be represented in a vector like form. A loop is not necessarily slower, when you deal with nested structs and cells.

Iniciar sesión para comentar.

Más respuestas (0)

Productos

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by