allow equals comparison of BigDecimals

[hughlunnon]

currently you can't easily compare bigdecimals for equality in kotest (ie : BigDecimal("1.0") shouldBe BigDecimal("1") passing)

if we add the following

infix fun BigDecimal?.shouldEqual(other: BigDecimal?) = this?.compareTo(other) shouldBe 0

to the BigDecimal comparitors file, it'll make it easier

happy to raise a PR

[sksamuel]

We could also add an Eq instance for BigDecimal to make it work? then BigDecimal("1.0") shouldBe BigDecimal("1") would work which I think is more intuitive ?

[hughlunnon]

it is more intuitive but those two bigdecimals aren't equal according to the bigdecimal spec (BigDecimal("1.0")==BigDecimal("1") is false) due to the number of decimal spots. Changing "shouldBe" to return true in that circumstance might be undesirable for some folks?

[Kantis]

BigDecimal would have the folloring assertions then, given this suggestion

• shouldBe
• shouldEqual
• shouldBeEqual

I think it might be a bit confusing to know which to pick. Perhaps naming the new one shouldHaveSameValue would set it apart sufficiently?

[LeoColman]

We do have matchers for compareTo

• shouldBeEqualComparingTo

Would this solve the issue for different scales? BigDecimal returns 0 for BigDecimal("1.0").compareTo(BigDecimal("1")

So one could just use

BigDecimal("1.005") shouldBeEqualComparingTo BigDecimal("1.0050000")

[hughlunnon]

We do have matchers for compareTo

• shouldBeEqualComparingTo

Would this solve the issue for different scales? BigDecimal returns 0 for BigDecimal("1.0").compareTo(BigDecimal("1")

So one could just use

BigDecimal("1.005") shouldBeEqualComparingTo BigDecimal("1.0050000")

ah yes that does work, but I think (given a few of us weren't aware of it) - its not very discoverable. A function alias within the io.kotest.matchers.bigdecimal package would be nice to aid discovery?

[LeoColman]

I think so, yes!

Now what to call it is tough. Any idea?

[sksamuel]

shouldBeEqualWithoutScale? as a match in the bigdecimal package ?

[LeoColman]

shouldBeSameValue?

[sksamuel]

Value is a word that could be used either way. Is 5 and 5.0 the same value or not the same value in general?

[sksamuel]

@LeoColman what's your final thoughts before I implement it?

[LeoColman]

I'm trying to think if mathematically we have a term that means 'this number represents the same value as this one, in totality'

And I can only think of shouldBeNumericallyEquivalentTo or shouldBeMathematicallyEquivalentTo

[LeoColman]

So 15.05 shouldBeNumericallyEquivalentTo 15.050000

[sksamuel]

How come you don't like the word scale? Since that's what big decimal itself uses. https://medium.com/beingabetterdeveloper/comparing-bigdecimal-s-with-different-scales-2901bf26538f

[LeoColman]

Like shouldBeIgnoringScale or the one you suggested?

Nothing against it at all 😁 go for it

[sksamuel]

shouldBeIgnoringScale or shouldBeEqualIgnoringScale What do you think ?

[LeoColman]

shouldBeEqualIgnoringScale sounds better

[sksamuel]

@LeoColman https://github.com/kotest/kotest/pull/3924