Arithmetic operation bug

[Manoj Chourasia]

v1="128.015"
puts((v1.to_f*1000).to_i).to_s // giving 128014 instead of 128015

Why the the result is rounding up to 128014 but the correct result
should be 128015.

 

[Mirko Viviani]

the problem seems to be in to_i

v1=3D"128.015"
puts((v1.to_f*1000).to_i).to_s // giving 128014 instead of 128015

Why the the result is rounding up to 128014 but the correct result
should be 128015.

--=20
-- Mirko Viviani --
GPG-PGP Public Key: 0xE4E8FAB1
Fingerprint: 14D3 A373 E926 7737 DF32 502B A4C4 1CE2 E4E8 FAB1
***********************************************
"=93Machines take me by surprise with great frequency.=94 A. Turing

 

[Chris Hulan]

v1="128.015"
puts((v1.to_f*1000).to_i).to_s // giving 128014 instead of 128015

Why the the result is rounding up to 128014 but the correct result
should be 128015.

Floats are tricky:
(120..130).each{|x| puts x,((x+0.015)*1000).to_i}

Have a look at "What Every Computer Scientist Should Know About
Floating-Point Arithmetic" (http://docs.sun.com/source/806-3568/
ncg_goldberg.html)

 

[Stefano Crocco]

|v1=3D"128.015"
|puts((v1.to_f*1000).to_i).to_s // giving 128014 instead of 128015
|
|Why the the result is rounding up to 128014 but the correct result
|should be 128015.

I'm not an expert, but I think because when you write 128.015, the number y=
ou=20
actually get is 128.0149999999999863575... (because 128.015 can't be=20
represented as a number with a finite number of decimal digits in binary, j=
ust=20
like, for example, 1/17 can't be repsented with a finite decimal number in=
=20
decimal). You usually don't see this because by default only a small number=
of=20
decimal digits are shown, taking rounding into account. You can obtain a=20
greater number of digits using, for example, format:

puts format("%.20f", 128.015)
=3D> 128.01499999999998635758

Now, when you multiply the number by 1000, what you actually get is:=20
128014.99999999998544808477. And, since Float#to_i truncates the number (it=
=20
doesn't round it), you get the number 128014.

I hope this helps

Stefano