Mathf.Max for absolute values?

[lone tapir]

Hi! sort of a short question, but how can you do Mathf.Max that checks the absolute values?

i.e max(-5,3) returns -5, max(6,9) returns 9, max(-2, 1) returns -2

[lavish lion]

Use an if statement?

[carmine stone]

There is no built in method for this but you can trivially write your own

[granite sage]

so you want the max function to return the input with higher magnitude?
simply use, Mathf.Max(Mathf.Abs(input1),Mathf.Abs(input2))

[carmine stone]

That's not what they asked. They want it to return the original number

As in returns the negative number if its absolute value is greater

[granite sage]

oh i see

[carmine stone]

f(-5, 3) should return -5, not 5

[granite sage]

we can easily achieve that with some if statements tho

[carmine stone]

indeed

[granite sage]

doubt
what do you want to do if max(-5,5)

[lone tapir]

yea i just used linq

i was trying to think of if statements like

but id have to convert them all to abs then mathf.max and somehow find the original one in the array

[carmine stone]

Ew

That's horrible

[lone tapir]

yea

should rewrite it tbh

[carmine stone]

Just write your own Max function it'd be like

2 lines of code

Hell you could one line it

without LINQ

[lone tapir]

how the hell would you do this in one line

youve got to loop through, check values and set a variable to the iterated one if it's bigger than the variable

[carmine stone]

what

Why would you need a loop

[lone tapir]

well we wanna get the biggest number in an array right?

thats how you figure out the biggest one, you need to check each one

[carmine stone]

You didn't mention an array in your question

[lone tapir]

i thought itd be a given

since max functions usually take arrays

[carmine stone]

No they don't

[lone tapir]

so anyways

[carmine stone]

They take just two arguments

That's Unity being stupid

[lone tapir]

hm

neat

[carmine stone]

But I will just give you a hint

first, are you working exclusively with floats? or do you need a solution that works for ints too

(or any other numeric type)

[lone tapir]

nvm just did it in one line

public static float SignedMax(float val1, float val2)
{
    return Mathf.Abs(val1) > Mathf.Abs(val2) ? val1 : val2;
}
public static float SignedMin(float val1, float val2)
{
    return Mathf.Abs(val1) < Mathf.Abs(val2) ? val1 : val2;
}

[carmine stone]

That won't necessarily give you the correct answer

[lone tapir]

i imagine it would

[carmine stone]

It returns a different value based purely on argument order

Which is horrendously bad design

[lone tapir]

i think thats just a floating point thing because its literally making both of them absolute

abs 5 is 5

abs -5 is 5

[carmine stone]

It's not a floating point thing

[lone tapir]

they're both equal, and its got to return one

[carmine stone]

The hint: .NET already has this feature. Argument order doesn't matter

if one value is negative and the other is positive, but their abs values are the same, it returns the positive

because the positive is still the max

[lone tapir]

the only way 5 and 5 arent equal would be floating point stuff

[lone tapir]

oh neat

[carmine stone]

This API isn't available in Unity because Unity is stupid and outdated and old

But .NET is open source so you can just

Yoink the code for it

https://source.dot.net/#System.Private.CoreLib/src/libraries/System.Private.CoreLib/src/System/MathF.cs,252

Go nuts

[lone tapir]

public static float MaxMagnitude(float val1, float val2)
{
    float abs1 = Mathf.Abs(val1);
    float abs2 = Mathf.Abs(val2);

    //1 more than 2, or 1 is so big its not even a number, return 1
    if ((abs1 > abs2) || float.IsNaN(abs1))
    {
        return val1;
    }

    //If both equal, return the regular biggest one
    if (abs1 == abs2)
    {
        return Mathf.Max(val1, val2);
    }

    //Otherwise idfk dude we already compared greater and equal cases just fucking return something
    //I love using vulgar language in code comments it reminds me of valve employees losing their mind writing TF2 spaghetti
    return val2;
}

teehee

[lavish lion]

So what I'm hearing is that you need to do

public static float SignedMax(float val1, float val2)
{
    return Mathf.Abs(val1) == Mathf.Abs(val2) ? Mathf.Max(val1, val2) : ((Mathf.Abs(val1) > Mathf.Abs(val2) || float.IsNan(Mathf.Abs(val1)) ? val1 : val2);
}

[carmine stone]

Checks out

[lone tapir]

EW

[mellow pike]

bro

😭

that is hurting my eyes

I guess that the other options is multiple lines