Why does python make mistake in basic arithmetic calculations?

Question

Try to calculate the following on python shell:

>>> 5.3-2.7
2.5999999999999996
>>> 2.8-2.7
0.09999999999999964
>>> 4.7-2.8
1.9000000000000004
>>> 4.3-2.5
1.7999999999999998
>>>

Why does it happen? How can I prevent this problem from making bugs in my code?

Please read this: https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html — Dair, Nov 14 '14 at 05:21
[What Every Programmer Should Know About Floating Point Arithmetic](http://floating-point-gui.de/) — bereal, Nov 14 '14 at 05:21
This is not an error but a limitation of floating point arithmetic. Google floating point arithmetic for many articles that can explain what's going on. — , Nov 14 '14 at 05:21
Worth taking a look at [the Python tutorial on floating point](https://docs.python.org/3.4/tutorial/floatingpoint.html) as well. — Sean Vieira, Nov 14 '14 at 05:22
Basically, floating point numbers are not completely accurate. — , Nov 14 '14 at 05:56

score -2 · Answer 1 · answered Nov 14 '14 at 05:21

-2

It's not a bug, but you can do this.

>>> round(5.3 - 2.7, 1)
>>> 2.6

answered Nov 14 '14 at 05:21

Aesthete

Meh.. Downvote me all you want, maybe someone is interested in actually seeing the result they expect. – Aesthete Nov 14 '14 at 05:23
Maybe, someone is interested in actually understanding what's happening, or how to deal with numbers other than in the example, like with more than one digit after the period. – bereal Nov 14 '14 at 05:27

1 Answers1