How do I define a structure that has an element which is an array?

How do I define a structure that has an element which is an array and for which there are sub elements. I have composed a simple example below of what I want to do along with a failed attempt at the needed assert statements.

function y = payroll(personnel) 
%#codegen

% Specify the class of the input as struct.
assert(isstruct(personnel));

% Specify the class and size of the fields r and i
% in the order in which you defined them.
assert(isa(personnel.num_employees,'int32'));
assert(isa(personnel.employee,'int32'));
assert(all(size(personnel.employee) == [5 1]));
assert(isa(personnel.employee(1).salary,'int32'));

for i = personnel.num_employees
   total_payroll = total_payroll + personnel.name(i).salary;

end

y = total_payroll;

end

[Arnab De]

Do you just want to run the MATLAB code or do you want to generate C code from it?

[Wayne Prather]

My purpose is to generate C code using Matlab Coder. I get an error related to my failure to properly define the employee element of my structure since it is an array and has a sub element uder it.

[Sean de Wolski]

Good thread +1

Answer 1

[Fred Smith]

This line seems wrong:

 assert(isa(personnel.employee,'int32'));

In your code employee is supposed to be a struct with sub-field name and salary? Once you make it a struct, you then also need to set its class and size of the fields name and salary.

-Fred

[Wayne Prather]

It finally works. Apparently you have to define the arrayed sub elements in the same way that you define the main structure. The following Matlab code works and succesfully produces the C++ code that follows.

Matlab code:

function y = payroll(personnel) 
%#codegen

% Specify the class of the input as struct.
assert(isstruct(personnel));

% Specify the class and size of the fields r and i
% in the order in which you defined them.
assert(isa(personnel.num_employees,'int32'));
assert(isstruct(personnel.employee));
assert(all(size(personnel.employee) == [5 1]));
assert(isa(personnel.employee(1).salary,'double'));

total_payroll = 0;

for i = 1 : personnel.num_employees
   total_payroll = total_payroll + personnel.employee(i).salary;   
end

y = total_payroll;

end

Matlab Coder produced C++ code:

real_T payroll(const b_struct_T personnel)
{
  real_T y;
  int32_T i;

    y = 0.0;
    for (i = 1; i <= personnel.num_employees; i++) {
      y += personnel.employee[i - 1].salary;

    }

return y;
}

Answer 2

[John Petersen]

1. Find out which assertion fails. The code looks fine.

2. Replace the 'for' loop with

y = sum([personnel.name.salary]);

Answer 3

[Jan]

This will crash, if the personnel.employee has more than 2 dimensions:

all(size(personnel.employee) == [5 1])

Better:

isequal(size(personnel.employee), [5 1])

Answer 4

[Sean de Wolski]

   total_payroll = total_payroll + personnel.name(i).salary;

total_payroll is never defined before being used for the first time!

[Wayne Prather]

Thanks, you are correct. I forgot to initialize it. The code is just a made up simplistic example that addresses the problem I am having in my real code in regards to defining a structure with elements which are arrays for the C++ code generation.