conversion error from double in c++ to long or float in kdb+

Community Support
1

I am trying to convert some double types in c++ to longs or floats in kdb+

using kj and KJ to convert to kdb+ the following nos in c++

1068080561

-639597367

-1108929061

-1115727465

1072032170

get converted to 

47524255114672j and 4 other nos differing only in the last few digits

using kf and KF they get converted to 

2.328563e-310  and nos differing only in the last few digits.   Can you give some help here?

Many thanks 

2

please can you post code that exhibits the issue?

The following works for me. What are you doing differently?

$ more guy.cpp
#include"k.h"

extern “C” K f(K x){
  K y;
  J i;
  if(KF==xt){
    y=ktn(KJ,xn);
    for(i=0;i<xn;i++)
      kJ(y)[i]=kF(x)[i];
  }
  else if(-KF==xt)
    y=kj(xf);
  else
    y=krr((S)“type”);
  return y;
}
   
$ more guy.q
f:($"./guy")2:(f;1)
a:1068080561 -639597367 -1108929061 -1115727465 1072032170f
f a
f each a
$ g++ -bundle -undefined dynamic_lookup guy.cpp -DKXVER=3 -fPIC -o2 -o guy.so -m32
$ ./q guy.q
KDB+ 3.3t 2015.05.27 Copyright (C) 1993-2015 Kx Systems
m32/ 4()core ..

1068080561 -639597367 -1108929061 -1115727465 1072032170
1068080561 -639597367 -1108929061 -1115727465 1072032170
q)\

likewise the following works for me

#include"k.h"

extern “C” K f(K x){
  double d={1068080561,-639597367,-1108929061,-1115727465,1072032170};
  K y=ktn(KF,5);
  J i;
  for(i=0;i<y->n;i++)
    kF(y)[i]=d[i];
  return y;
}

$ ./q
KDB+ 3.3t 2015.05.27 Copyright (C) 1993-2015 Kx Systems
m32/ 4()core
q)\P 0
f:($"./guy")2:(f;1)q)
q)f
1068080561 -639597367 -1108929061 -1115727465 1072032170f
q)\