NoComparisonOnFloats.qhelp

<!DOCTYPE qhelp PUBLIC
  "-//Semmle//qhelp//EN"
  "qhelp.dtd">
<qhelp>


<overview>
<p>
Equality tests on floating point values
may lead to unexpected results because of arithmetic imprecision.
For example, the expression <code>23.42f==23.42</code>
evaluates to <code>false</code>.
</p>

</overview>
<recommendation>

<p>Instead of testing for <i>exact equality</i> between floating point values, check that the 
difference between the values is within an appropriate error margin.</p>

<p>Alternatively, if you do not want any inaccuracy when testing for equality, use one of the 
following instead of floating point values:</p>
<ul>
    <li><code>BigDecimal</code> class. This can store decimal values with higher precision.</li>
    <li><code>long</code> type. Because this is an integer type, you must convert any decimal values
    to whole values. For example, represent $1.43 as 143 cents.</li>
</ul>

</recommendation>
<example>
<p>In the following example, <code>(0.1 + 0.2) == 0.3</code> evaluates to <code>false</code>, even
though you would expect it to evaluate to <code>true</code>. This is because of the imprecision of 
floating point data types.</p>
<sample src="NoComparisonOnFloats.java" />

<p>In the following improved example, the test for equality is performed by calculating the difference
between the two values, and checking if the difference is within the error margin, <code>EPSILON</code>.</p>
<sample src="NoComparisonOnFloatsGood.java" />

</example>
<references>


<li>
J. Bloch, <em>Effective Java (second edition)</em>, Item 48. Addison-Wesley, 2008.
</li>
<li>
Numerical Computation Guide: 
(<a href="http://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html">What Every Computer Scientist Should Know About Floating-Point Arithmetic</a>).
</li>


</references>
</qhelp>