Can anyone tell me why the following statement evaluates to false?
bool myBoolean = .6 + .3 + .1 == .1 + .3 + .6; // false
It does the same thing in Javascript and C++.
Asked
Active
Viewed 151 times
5
-
1Floating point arithmetic? – Soner Gönül Apr 22 '15 at 08:19
-
As a note, testing for float equality after any arithmetic is usually a bad idea – Sebi Apr 22 '15 at 08:21
-
1Use the right type `decimal`: `bool myBoolean = .6m + .3m + .1m == .1m + .3m + .6m;` – Tim Schmelter Apr 22 '15 at 08:21
-
Never trust floating point equality comparisons. You can simplify your example as bool myBoolean = .1 == .1; – Callum Bradbury Apr 22 '15 at 08:21
-
@CallumBradbury: i cannot reproduce what you say. `bool myBoolean = .1 == .1;` returns true for me. That's confusing because the highly upvoted duplicate-question also states that. – Tim Schmelter Apr 22 '15 at 08:32
2 Answers
7
Nb. I'm answering this question since it is also tagged C++
Due to float representation errors, in C++ the above yields false since the two float numbers aren't perfectly equal.
E.g. the internal representation for 0.1 is close to that value, but not exactly it
The same holds for C#.
Allow me to link a famous document that (imho) every programmer dealing with floating point arithmetic should read: What Every Computer Scientist Should Know About Floating-Point Arithmetic
Marco A.
- 43,032
- 26
- 132
- 246
0
]Never trust the Floating-Point Arithmetic. During the conversion you can have a loss of precison. That's why the result is false. I suggest u try using the decimal type which stores numbers in decimal notation. Thus your numbers will be representable precisely.
MRebai
- 5,344
- 3
- 33
- 52