double -> integer

New comment
 
Hello :), I need help with my function, I need to convert double to integer.
Here's example

0.4 -> 4
0.0023 -> 23
0.01004 -> 1004

and here's my function:

int cele_cislo(double n)
{
   int ret;
   double v1=0;
   double v2=n;
   double vynasobene=n;
   int i=0;
   
   while(v1!=v2)
   {
      vynasobene *= 10;
      
      v1 = vynasobene;
      
      ret = vynasobene;
      v2 = ret;
  
      i++;
      if(i>15) break;
      if(v1==v2) break;
   }
   
   return(ret);
}

it works fine when I test it, but when I use it in my EA, sometimes it returns strange values, or do you have any better idea how to do this? Thx :)
 
buksy:
Hello :), I need help with my function, I need to convert double to integer.
Here's example
[....]
it works fine when I test it, but when I use it in my EA, sometimes it returns strange values, or do you have any better idea how to do this? Thx :)

You are dealing with the "imprecise" nature of double arithmetic which has been discussed many times on this forum, for example in https://www.mql5.com/en/forum/116326. With at least one of your examples, there is the additional issue that MQ4 can display up to 8 decimal places using DoubleToStr(), but uses greater precision than this internally.

For example, when you multiply a double value of 0.01004 by 10, the values which are actually getting held in the double variable are as follows. MQ4 will show the final value as 1004 because it can only display up to 8DP.

0.100400000000000
1.004000000000000
10.039999999999999
100.399999999999990
1003.999999999999900

Your only option is to specify the amount of precision you want to use in your calculation, using the NormalizeDouble() function. For example, with the following function, cele_cislo(0.01004) will return 1004.

int cele_cislo(double x, int MaxPrecisionDP = 5)
{
   x = NormalizeDouble(x, MaxPrecisionDP);
   while (x - MathFloor(x) != 0) {
      x = NormalizeDouble(x * 10, MaxPrecisionDP);
   }
   return (MathFloor(x));
}
 
Thank you :)
New comment