reading a binary file of class with struct as member

[cdf]

I would like to read a binary file that has a number of class P objects, and one of the members is an array of struct C objects:

class P
{
protected:
double n[5];
C Cs[5];
};

struct C
{
int S;
char Name[20];
};

and I have no idea how to do this, i went thru some examples and they are for reading videos/images...
i know how to get the number P objects, but i dont know how to construct them using the data from the file...

[Stewbond]

If it contains only one C struct, do this:

// read a file into memory
#include <iostream>
#include <fstream>
using namespace std;

struct C
{
  int S;
  char Name[20]
};

int main () {
  C InputStruct;

  ifstream is;
  is.open ("test.txt", ios::binary );

  // read data as a block:
  is.read ( &InputStruct, sizeof(InputStruct) );
  is.close();

  return 0;
}

See
http://cplusplus.com/reference/iostream/istream/read/

[cdf]

i know how to do this, but the problem is reading class objects - are the objects already "constructed" inside the file, and all i do is assign them to an array of objets - or does the binary file contain only data, so i get these parameters and construct the objects?

and also i need to create the array of C structures within each P

if i have P temp;
is it possible that all i need to do is use seekg for 0 to number of P objects, each time using f.read ( &temp, sizeof(temp) ) and then copying temp into an array of P objets, while of course overloading the assignment operator so that any pointers dont point to the actual text file but to newly allocated memory where the members are copied?

[ne555]

You didn't show pointers there...
See http://www.cplusplus.com/forum/beginner/76800/

[Stewbond]

Ok, let's say the file uses this structure:
1st 4 bytes contain an integer with how many C structs are there,
2nd 4 bytes contain an integer with how many P classes are there,

Then do this:

// read a file into memory
#include <fstream>
using namespace std;

class P
{
protected:
  double n[5];
  C Cs[5];
};

struct C
{
  int S;
  char Name[20]
};

int main () {
  C* InputStructs;
  P* InputClasses;
  int NumClasses;
  int NumStructs;
  
  ifstream is;
  is.open ("test.txt", ios::binary );

  // Get the number of objects to read
  is.read ( (char*)&NumStructs, sizeof(NumStructs) );
  is.read ( (char*)&NumClasses, sizeof(NumClasses) );
  
  // Allocate the applicable amount of memory:
  InputStructs = new C[NumStructs];
  InputClasses = new P[NumClasses];
  
  // Option 1, read them one at a time:
  for (int i = 0; i < NumStructs; ++i)
    is.read ( (char*)&InputStructs[i], sizeof(InputStructs[i]) );
  for (int i = 0; i < NumClasses; ++i)
    is.read ( (char*)&InputClasses[i], sizeof(InputClasses[i]) );
  
  // Option 2, read large bulk 
  is.read ( (char*)InputStructs, sizeof(InputStructs[0]) * NumStructs);
  is.read ( (char*)InputClasses, sizeof(InputClasses[0]) * NumClasses);
  
  is.close();

  return 0;
}

Edit: I'm reading disch's post in ne555's link. It makes sense and would probably be a more professional solution.