To round or not to round… [Software]

posted by ElMaestro  – Denmark, 2019-07-20 21:59 (1731 d 19:45 ago) – Posting: # 20415
Views: 13,990

Hi Hötzi,

❝ I don’t know which kind of measurements you are comparing.


It stinks but there really is no good solution when you have to compare.
(294.1/1000) == 0.2941
[1] FALSE
(2941/10000) == 0.2941
[1] TRUE


One can read ISO standards and manuals about internal float representation (I tried andfell asleep), but at the end of the day if you need to compare you need to compare and hope for the best. Some numbers just cannot be represented accurately in binary.

If you accept an xyz fraction of error on the comparison then the solution is straightforward, like

Match = function (x,y, relative.tol)
{
 if (y<((1+relative.tol)*x))
  if (y>((1-relative.tol)*x))
    return(T)
 return(F)
}

Match(6, 6.001, 0.01)


Ugly as hell :vomit:, and will not work well if you get into extreme binary representations.

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
22,984 posts in 4,822 threads, 1,651 registered users;
54 visitors (0 registered, 54 guests [including 4 identified bots]).
Forum time: 17:45 CEST (Europe/Vienna)

You can’t fix by analysis
what you bungled by design.    Richard J. Light, Judith D. Singer, John B. Willett

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5