Python float equality weirdness

Question

Seeing some unexpected behavior with Python tonight. Why is the following printing out 'not equal'?!

num = 1.00
num -= .95
nickel = .05

if nickel != num:
    print 'not equal'
else:
    print 'equal'

Great stuff guys. Been around for a while and never ran into it. Thanks. — Benjamin Powers, May 12 '12 at 04:22

score 6 · Accepted Answer · answered May 12 '12 at 03:32

6

>>> num = 1.00
>>> num
1.0
>>> num -= 0.95
>>> num
0.050000000000000044
>>> nickel = .05
>>> nickel
0.05

answered May 12 '12 at 03:32

voithos

jgritty · Answer 2 · 2012-05-12T03:46:53.947

2

You might find the decimal module useful.

>>> TWOPLACES = Decimal(10) ** -2
>>> Decimal(1).quantize(TWOPLACES)-Decimal(0.95).quantize(TWOPLACES) == Decimal(0.05).quantize(TWOPLACES)
True

Or, alternatively:

import decimal
decimal.getcontext().prec = 2
decimal.Decimal(1.00) - decimal.Decimal(0.95)

I inferred from your naming of the nickel variable that you were thinking about money. Obviously, floating point is the wrong Type for that.

edited May 12 '12 at 03:46

answered May 12 '12 at 03:32

jgritty

I disagree, it's the same problem. `>>> Decimal(.05) == Decimal(1.00 - .95) >>> False` – jb. May 12 '12 at 03:35
2

`Decimal('1') - Decimal('.95') == Decimal('.05')` or `Decimal(str(1.00 - .95)) == Decimal('.05')` – jamylak May 12 '12 at 03:39
2

@jb. It's the same problem because you've converted the result of a floating point operation to a decimal. You need to have the operation done on decimals. You also have to be careful assigning decimals from floats. Note the difference between `Decimal(0.1)` and `Decimal('0.1')` – Paul S May 13 '12 at 23:53

score 0 · Answer 3 · answered May 12 '12 at 03:33

0

This is a common floating point problem with computers. It has to do with how the computer stores floating point numbers. I would suggest giving What Every Computer Scientist Should Know About Floating-Point Arithmetic a quick read through.

answered May 12 '12 at 03:33

jb.

3 Answers3