Float precision and float equality
in [Python]
|
From: sturlamolden on 5 Dec 2009 19:31 On 5 Des, 16:37, Anton81 <gerenu...(a)googlemail.com> wrote: > I'd like to do calculations with floats and at some point equality of > two number will be checked. > What is the best way to make sure that equality of floats will be > detected, where I assume that mismatches beyond a certain point are > due to truncation errors? isequal = lambda x,y : abs(x-y) < eps where eps is the truncation error.
From: Tim Roberts on 6 Dec 2009 02:42 Raymond Hettinger <python(a)rcn.com> wrote: > > if not round(x - y, 6): ... That's a dangerous suggestion. It only works if x and y happen to be roughly in the range of integers. For example, here x and y are within roundoff error of each other, but round doesn't know it: >>> x=1e32 >>> y=x+1e16 >>> x-y -18014398509481984.0 >>> round(x-y,6) -18014398509481984.0 It fails in the other direction when the numbers are small: >>> x=.0000000123 >>> y=.0000000234 >>> x-y -1.1100000000000002e-008 >>> round(x-y,6) 0.0 Mark's solution is the generically correct one, which takes into account the rough range of the values. -- Tim Roberts, timr(a)probo.com Providenza & Boekelheide, Inc.
From: Raymond Hettinger on 6 Dec 2009 04:12 On Dec 5, 11:42 pm, Tim Roberts <t...(a)probo.com> wrote: > Raymond Hettinger <pyt...(a)rcn.com> wrote: > > > if not round(x - y, 6): ... > > That's a dangerous suggestion. It only works if x and y happen to be > roughly in the range of integers. Right. Using abs(x-y) < eps is the way to go. Raymond
From: dbd on 6 Dec 2009 14:23 On Dec 6, 1:12 am, Raymond Hettinger <pyt...(a)rcn.com> wrote: > On Dec 5, 11:42 pm, Tim Roberts <t...(a)probo.com> wrote: > > > Raymond Hettinger <pyt...(a)rcn.com> wrote: > > > > if not round(x - y, 6): ... > > > That's a dangerous suggestion. It only works if x and y happen to be > > roughly in the range of integers. ..> ..> Right. Using abs(x-y) < eps is the way to go. ..> ..> Raymond This only works when abs(x) and abs(y) are larger that eps, but not too much larger. Mark's suggestion is longer, but it works. The downside is it requires you to think about the scale and accuracy of your application. Dale B. Dalrymple
From: Anton81 on 6 Dec 2009 14:34
I do some linear algebra and whenever the prefactor of a vector turns out to be zero, I want to remove it. I'd like to keep the system comfortable. So basically I should write a new class for numbers that has it's own __eq__ operator? Is there an existing module for that? |