I want to know how to convert values of float and double into binary format and push into vector of type uint8_t

Eg : float x  = 23.22;
     double z = 2.32232;

and store them into vector while serializing

vector<uint8_t> data.

and also convert back into original value while deserializing.

Is there any way to do this?

CodePudding user response：

If you just need to push them into vector and pop them (like a stack) you can do this:

void push( std::vector<uint8_t> &v, float f )
{
   auto offs = v.size();
   v.resize( offs   sizeof( float ) );
   std::memcpy( v.data()   offs, &float, sizeof( float ) );
}

float popFloat( std::vector<uint8_t> &v )
{
    float f = 0;
    if( v.size() >= sizeof( float ) ) {
        auto offs = v.size() - sizeof( float );
        std::memcpy( &f, v.data()   offs, sizeof( float ) );
        v.resize( offs );
    }
    return f;
}

Note this would store them in not portable format, but should work for storing/reading them to/from file on the same hardware.

You may rewrite those 2 functions as template and it will work with all integral and floating point types ( short int long double etc )