float point issue in python list

Question

I am writing a piece of test code that add two float point int each other. I can get accuracy result,but the extra number of points appear when I insert the result in a python list, I have no idea what lead to this consequence. Please give me a hint!

My code:

   center = [120.688281,30.500036]
   coupon_list = []
   for i in range(5):
        seed_x = random.randint(1,100)
        print 'seed_x:'
        print seed_x
        random.seed(seed_x)
        rand_x = random.randrange(100,500,20)/float(100000)

        seed_y = random.randint(1,100)
        print 'seed_y:'
        print seed_y
        random.seed(seed_y)
        rand_y = random.randrange(100,500,20)/float(100000)
        print 'rand_x:'
        print rand_x
        print 'rand_y:'
        print rand_y
        print 'float convert:'
        x = center[0]+ rand_x
        y = center[1] + rand_y
        print 'x:'
        print x
        print 'y:'
        print y
        coupon = []
        coupon.append(x)
        coupon.append(y)
        print 'coupon:'
        print coupon
        coupon_list.append(coupon)
    print coupon_list

My result:

seed_x:
22
seed_y:
15
rand_x:
0.0048
rand_y:
0.0048
float convert:
x:
120.693081
y:
30.504836
coupon:
[120.693081, 30.504836]
seed_x:
2
seed_y:
95
rand_x:
0.0048
rand_y:
0.004
float convert:
x:
120.693081
y:
30.504036
coupon:
[120.693081, 30.504036000000003]
seed_x:
52
seed_y:
6
rand_x:
0.0048
rand_y:
0.004
float convert:
x:
120.693081
y:
30.504036
coupon:
[120.693081, 30.504036000000003]
seed_x:
83
seed_y:
86
rand_x:
0.0028
rand_y:
0.004
float convert:
x:
120.691081
y:
30.504036
coupon:
[120.691081, 30.504036000000003]
seed_x:
4
seed_y:
11
rand_x:
0.0018
rand_y:
0.0028
float convert:
x:
120.690081
y:
30.502836
coupon:
[120.690081, 30.502836000000002]
[[120.693081, 30.504836], [120.693081, 30.504036000000003], [120.693081, 30.504036000000003], [120.691081, 30.504036000000003], [120.690081, 30.502836000000002]]

There is no issue here. Objects in a list are represented by their repr() output. Your float object has that exact value, but the str() conversion limits the number of digits that are printed. — Martijn Pieters
– Martijn Pieters, Commented Nov 3, 2016 at 9:23
@ Martijin Pieters Why some results appear the extra number of points and some are not? If it is not a issue, all the result should print out the same number of points. — pipi
– pipi, Commented Nov 3, 2016 at 9:33
Floating point numbers use binary fractions to approximate values. Not all values can be represented exactly using binary fractions. You have some such numbers, so there is a very, very small deviation to make it fit. — Martijn Pieters
– Martijn Pieters, Commented Nov 3, 2016 at 9:36

Rolf of Saxony · Accepted Answer · 2016-11-03 14:06:44Z

2

If you can decide on how much precision you require, you could could always use something like this.

coupon.append(float(format(x,'.6f')))
coupon.append(float(format(y,'.6f')))

Which in this case gives 6 digits after the decimal point.
e.g.

>>> a=30.504036000000003
>>> print a
30.504036
>>> repr(a)
'30.504036000000003'
>>> coupon=[]
>>> coupon.append(a)
>>> coupon
[30.504036000000003]
>>> coupon.append(float(format(a,'.6f')))
>>> coupon
[30.504036000000003, 30.504036]
>>> coupon[0]+coupon[1]
61.008072
>>> type(coupon[0])
<type 'float'>
>>> type(coupon[1])
<type 'float'>

Edit: or as Mark pointed out in his comment you could instead use:

coupon.append(round(x,6))
coupon.append(round(y,6))

and get the same result.

edited Nov 3, 2016 at 14:06

answered Nov 3, 2016 at 9:53

Rolf of Saxony

22.6k5 gold badges43 silver badges61 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Mark Dickinson Over a year ago

float(format(a, '.6f')) is a rather complicated way of saying round(a, 6).

Rolf of Saxony Over a year ago

@MarkDickinson coupon.append(round(a,6)) Your right! Checked it and you get a float stored in coupon. I didn't think of it but that's the beauty of python, there's always more than one way to skin a cat.

Collectives™ on Stack Overflow

float point issue in python list

1 Answer 1

2 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Related