BigDecimal::of(0.00001)->getScale() is 6 about math

[BenMorel]

Hi, these discrepancies appear because you're using floating-point values:

echo $a = BigDecimal::of(0.00001); // 0.000010 (!)
echo $a->getUnscaledValue(); // 10
echo $a->getScale(); // 6

As advised in the README, you cannot trust a floating point value, and should always instantiate from an integer or string value:

echo $b = BigDecimal::of('0.00001'); // 0.00001
echo $b->getUnscaledValue(); // 1
echo $b->getScale(); // 5

Note that in this case, only the scale differs, the values are still equal:

var_export($a->isEqualTo($b)); // true

---

Where does the extra digit come from?

In this particular case, the reason is not a rounding error (a common issue with floating point numbers), but the fact that the value is converted to a string, then parsed:

echo (string) 0.00001; // 1.0E-5

The fact that it returns a number with value 10 and scale 6 is consistent with the fact that for the number 1.1E-5, one would expect value 11 and scale 6.

On the other hand, 1E-5 would return what you expect:

echo BigDecimal::of('1E-5')->getUnscaledValue(); // 1
echo BigDecimal::of('1E-5')->getScale(); // 5

In summary, no bug here, just use strings! 👍

[lx495707690]

@BenMorel Thank you for your reply！