Float to Int

Stream: Is there code for X?

[Marcus Rossel]

Is there any way to convert a Float to an adjacent Int? Rounding up, down, anything.

[Martin Dvořák]

@loogle (Float -> Int)

[loogle]

:shrug: nothing found

[Edward van de Meent]

you may be able to go via one of variousUInt types?

[Edward van de Meent]

alternatively, use Float.toRat0?

[Adam Topaz]

docs#Float.floor

[Adam Topaz]

Ha but that gives a float back

[Edward van de Meent]

surprising, right?

[Julian Berman]

What would it do to NaN if it didn't return a float

[Julian Berman]

(I guess in the original case you don't care, but yeah it seems like any "real" defined coercion needs to decide on what to do with infs an nans?)

[Edward van de Meent]

the toRat0 function i referenced returns 0, but there is also a version with codomain Option Rat, which returns Option.none in those cases.

[Mario Carneiro]

Using conversion to Rat is the best option if you want exact values. Unfortunately there is no Rat.round function yet...

[Eric Wieser]

Don't we have docs#round ?

[Mario Carneiro]

ah, it's not in batteries yet

[Mario Carneiro]

it's also not using ties-to-even

[Eric Wieser]

Nor is it documented as not being that

[Edward van de Meent]

as a last resort, you can try using Float.toRatParts, but that might be a bit more hands-on