Rounding numbers in Python

Question

a=0.005
print ('%.2f'%a)
b=90.005
print('%.2f'%b)
c=90.015
print('%.2f'%c)

Above code is written in python3 and the output is following:

 0.01
 90.00
 90.02

Is this any kind of computational error or m missing a point?

@mhd That doesn't explain why the first and last are rounded — OneCricketeer, Jun 01 '18 at 13:36
If I'd to guess, it's somewhat related to this https://stackoverflow.com/questions/588004/is-floating-point-math-broken — OneCricketeer, Jun 01 '18 at 13:38
I think this explains it: https://stackoverflow.com/questions/10825926/python-3-x-rounding-behavior — sjw, Jun 01 '18 at 13:41
Possible duplicate of [Python 3.x rounding behavior](https://stackoverflow.com/questions/10825926/python-3-x-rounding-behavior) — divibisan, Jun 01 '18 at 13:48
@thesilkworm that question answers my problem properly. Looking for a better answer — Shivendra Saxena, Jun 01 '18 at 13:48
What's your question? When you round a number to 2 digits, if the 3rd digit is 0-4 it is rounded down while if the 3rd digit is 5-9 it is rounded up. Your example is exactly what is expected when rounding numbers. Are you trying to do something different? — divibisan, Jun 01 '18 at 13:50

score 2 · Accepted Answer · answered Jun 01 '18 at 13:56

2

I think the problem here is the second one, which is

b = 90.005
print('%.2f' % b)

Now as @divibisan said, if the digit is between 0-4, it's rounded down, if it's between 5-9, it's rounded up. So why is 90.005 not rounded up ?

Because computers can not represent the floating points precisely. If you look at it closely you will see that 90.005 is represented as,

>>> print('%.50f' % b)
90.00499999999999545252649113535881042480468750000000

That's the reason it is being rounded down, while others behave normally.

>>> print ('%.50f'%a)
0.00500000000000000010408340855860842566471546888351
# 0.01

>>> print ('%.50f'%c)
90.01500000000000056843418860808014869689941406250000
# 90.02

answered Jun 01 '18 at 13:56

BcK

2,548
1
13
27

Yupp, i understand this, So whats the possible solution for this, I face this problem several times at HakerRank – Shivendra Saxena Jun 01 '18 at 13:59
You need to convert them to integers, do calculations and after that convert to floating points. – BcK Jun 01 '18 at 14:00
OK, means there is no efficient solution for it. – Shivendra Saxena Jun 01 '18 at 14:02
@ShivendraSaxena From the documentation: **For use cases which require exact decimal representation, try using the decimal module which implements decimal arithmetic suitable for accounting applications and high-precision applications.** – BcK Jun 01 '18 at 14:04
Can u provide me reference link of that doc – Shivendra Saxena Jun 01 '18 at 14:12
@Shivendra Saxena https://docs.python.org/3/tutorial/floatingpoint.html – BcK Jun 01 '18 at 14:19

Rounding numbers in Python

1 Answers1