``` assertEquals( BigDecimal("0...

# announcements

[Joel]

assertEquals(
            BigDecimal("0.08333333333333333"),
            (1.toBigDecimal() / 12.toBigDecimal())
        )

expected: <0.08333333333333333> but was: <0>

The bottom evaluates to 0 which is clearly incorrect. What is the rule here and how can I avoid this pitfall?

[Robert Jaros]

1.toBigDecimal().setScale(17) / 12.toBigDecimal()

[Joel]

I’d have to determine a scale beforehand?

[Robert Jaros]

generally you have to specify the scale of the result - that's what BigDecimal class is all about

1.toBigDecimal().divide(12.toBigDecimal(), 17, java.math.BigDecimal.ROUND_HALF_UP)

[Robert Jaros]

https://kotlinlang.org/api/latest/jvm/stdlib/kotlin/java.math.-big-decimal/div.html

[Joel]

Thank you @Robert Jaros

[Joel]

It looks like this is actually related to Kotlin:

@kotlin.internal.InlineOnly
public inline operator fun BigDecimal.div(other: BigDecimal): BigDecimal = this.divide(other, RoundingMode.HALF_EVEN)

The default divide operator uses half even rounding and determines scale from the operating decimals. If neither have a scale, it rounds to the nearest integer. You can override the division operator using:

private operator fun BigDecimal.div(other: BigDecimal) = divide(other, MathContext.DECIMAL128)

which changes the behavior to compute the decimal positions.