in [Python]

From: Patrick Maupin on 1 Apr 2010 22:01 On Apr 1, 7:49 pm, Tim Chase <python.l... (a)tim.thechases.com> wrote:> David Robinow wrote: > > $ python -c "print 1/2 * 1/2" > > 0 > > > But that's not what I learned in grade school. > > (Maybe I should upgrade to 3.1?) > > That's because you need to promote one of them to a float so you > get a floating-point result: > > >>> 1/2 * 1/2 > 0 > >>> 1/2 * 1/2.0 > 0.0 > > Oh...wait ;-) > > -tkc Hmmm, I think I'm starting to see why we need math.fsum() to take care of those rounding errors...
From: Steven D'Aprano on 1 Apr 2010 22:38 On Thu, 01 Apr 2010 19:55:27 -0400, David Robinow wrote: >>> superpollo wrote: >>> > how much is one half times one half? [...] > Well, my python says: > > $ python -c "print 1/2 * 1/2" > 0 > > But that's not what I learned in grade school. > (Maybe I should upgrade to 3.1?) Python 2.x defaults to integer division, a design error which has been rectified in 3.x. One can do any of these: [steve (a)sylar ~]$ python3.1 -c "print(1/2 * 1/2)"0.25 [steve (a)sylar ~]$ python2.6 -c "from __future__ import division; print 1/2 * 1/2" 0.25 [steve (a)sylar ~]$ python2.6 -Q new -c "print 1/2 * 1/2"0.25 [steve (a)sylar ~]$ python2.6 -c "print 0.5 * 0.5"0.25 and probably many others as well. -- Steven
From: Steven D'Aprano on 1 Apr 2010 22:44 On Thu, 01 Apr 2010 19:49:43 -0500, Tim Chase wrote: > David Robinow wrote: >> $ python -c "print 1/2 * 1/2" >> 0 >> >> But that's not what I learned in grade school. >> (Maybe I should upgrade to 3.1?) > > That's because you need to promote one of them to a float so you get a > floating-point result: > > >>> 1/2 * 1/2 > 0 > >>> 1/2 * 1/2.0 > 0.0 > > Oh...wait ;-) Tim, I'm sure you know the answer to this, but for the benefit of the Original Poster, the problem is that you need to promote *both* divisions to floating point. Otherwise one of them will give int 0, which gives 0.0 when multiplied by 0.5. >>> 1.0/2 * 1/2.0 0.25 If you want an exact result when multiplying arbitrary fractions, you need to avoid floats and decimals and use Fractions: >>> Fraction(1, 2)**2 Fraction(1, 4) -- Steven
From: Lie Ryan on 1 Apr 2010 22:50 On 04/02/10 13:01, Patrick Maupin wrote: > On Apr 1, 7:49 pm, Tim Chase <python.l... (a)tim.thechases.com> wrote:>> David Robinow wrote: >>> $ python -c "print 1/2 * 1/2" >>> 0 >> >>> But that's not what I learned in grade school. >>> (Maybe I should upgrade to 3.1?) >> >> That's because you need to promote one of them to a float so you >> get a floating-point result: >> >> >>> 1/2 * 1/2 >> 0 >> >>> 1/2 * 1/2.0 >> 0.0 >> >> Oh...wait ;-) >> >> -tkc > > Hmmm, I think I'm starting to see why we need math.fsum() to take care > of those rounding errors... hmm? >>> import math >>> math.fsum([1/2, 1/2]) 0.0 it doesn't appear to take care of those rounding errors, not in this case at least.
From: Patrick Maupin on 1 Apr 2010 22:58
On Apr 1, 9:50 pm, Lie Ryan <lie.1... (a)gmail.com> wrote:> On 04/02/10 13:01, Patrick Maupin wrote: > > > > > On Apr 1, 7:49 pm, Tim Chase <python.l... (a)tim.thechases.com> wrote:> >> David Robinow wrote: > >>> $ python -c "print 1/2 * 1/2" > >>> 0 > > >>> But that's not what I learned in grade school. > >>> (Maybe I should upgrade to 3.1?) > > >> That's because you need to promote one of them to a float so you > >> get a floating-point result: > > >> >>> 1/2 * 1/2 > >> 0 > >> >>> 1/2 * 1/2.0 > >> 0.0 > > >> Oh...wait ;-) > > >> -tkc > > > Hmmm, I think I'm starting to see why we need math.fsum() to take care > > of those rounding errors... > > hmm? > > >>> import math > >>> math.fsum([1/2, 1/2]) > > 0.0 > > it doesn't appear to take care of those rounding errors, not in this > case at least. you're right! I mis-read the problem. What we REALLY need is a good math.fmul() ;-) |