How to output floating point numbers with a custom output format in C++?

Question

The problem is that I want to output Mathematica compatible floating point numbers. The only difference with the standard IOStream or printf output format is that the exponential e is replaced by *^:

Standard C/C++ output format: 1.23e-4
Mathematica format: 1.23*^-4

Is there a way to manipulate streams to achieve this effect? My original idea was just to use istringstream and dump it to a string and then replace all the e's. I would also be okay if someone posted code to parse through the bits of the floating point number and output it directly (i.e. a printf("%e") replacement).

Answer 1

Your idea should work out easily:

std::string FloatToStringMathematica(float x) {
  std::stringstream a;
  a << x;
  std::string u = a.str();
  size_t pos = u.find("e");
  if(pos != std::string::npos) {
    u.replace(p, 1, "*^"); 
  }
  return u;
}

Answer 2

printf with %f shouldn't use scientific notation at all. Is there some reason you must use this notation and not simply output a long enough string to fit all significant figures?