Q&A's, tips, howto's
HPW
Posts: 1390 Joined: Thu Sep 26, 2002 9:15 am
Location: Germany
Contact:
Post
by HPW » Wed Mar 20, 2019 10:34 am
Hello,
I get this unexpected result in round:
Code: Select all
> (round 8.3 -1)
8.300000000000001
> (round 8.30 -1)
8.300000000000001
> (round 8.430 -1)
8.4
> (round 8.4 -1)
8.4
> (round 8.2 -1)
8.199999999999999
>
The doc contains not this behaviour.
Regards
Hans-Peter
Hans-Peter
ralph.ronnquist
Posts: 228 Joined: Mon Jun 02, 2014 1:40 am
Location: Melbourne, Australia
Post
by ralph.ronnquist » Wed Mar 20, 2019 11:46 am
If you don't mind, I would point you at this fairly in-depth discussion about "representable numbers"
https://en.wikipedia.org/wiki/Floating- ... d_rounding
which I think is the issue at hand here.
In short, the numbers you see printed are the nearest representable numbers to the numbers you asked for, given that the rounding is done wrt a decimal base whilst the actual values are held in binary.
Lutz
Posts: 5289 Joined: Thu Sep 26, 2002 4:45 pm
Location: Pasadena, California
Contact:
Post
by Lutz » Wed Mar 20, 2019 2:53 pm
If you are concerned about the visual representation use format:
Code: Select all
> > (round 8.300 -1)
8.300000000000001
> (format "%2.1f" 8.300)
"8.3"
>
HPW
Posts: 1390 Joined: Thu Sep 26, 2002 9:15 am
Location: Germany
Contact:
Post
by HPW » Wed Mar 20, 2019 4:37 pm
Hello,
Thanks for the advices.
I am generating nc-code for a nc-drilling mashine.
Format will do the Job.
Regards
Hans-Peter
Hans-Peter