🔒 🔒 zero rounding of float | Python | Page 1

stray frost Jan 5, 2025, 9:34 PM

#

rounding
a = round(-0.0001,2)
print(a)
-0.0

if there a way take advantage of this and test if it really is a round from a negative number with standard math or logical ? ....
maybe converting it in to string and analysing it...

mental hawkBOT Jan 5, 2025, 9:34 PM

#

stray frost rounding a = round(-0.0001,2) print(a) -0.0 if there a way take advantage of t...

#

@stray frost

Python help channel opened

Remember to:

Ask your Python question, not if you can ask or if there's an expert who can help.
Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

deft granite Jan 5, 2025, 9:37 PM

#

I don't really understand your question, what do you mean? Test if it's close to something?

grizzled drift Jan 5, 2025, 9:37 PM

#

Not sure about what you are trying to achieve there, what are you trying to do exactly ?

#

Damn, mods are fast

stray frost Jan 5, 2025, 9:38 PM

#

in some case rouding from negative number give zerro doesnt have the same meaning from rounding a positive number to zero.

#

it seems float can give -0.0 from a negative number

#

but -0.0 == 0.0

#

if I used =

deft granite Jan 5, 2025, 9:40 PM

#

So you want rounding, but negative zero turned to positive zero?

stray frost Jan 5, 2025, 9:41 PM

#

? in some case rouding from negative number give zerro doesnt have the same meaning from rounding a positive number to zero.

#

I want yo detct the -0.0

deft granite Jan 5, 2025, 9:41 PM

#

Ah, and you think you can't because 0.0 == -0.0

stray frost Jan 5, 2025, 9:41 PM

#

and the 0.0 separatly

#

-0.0 < 0 False

grizzled drift Jan 5, 2025, 9:42 PM

#

The real question is why you would like to detect it thought, is it just curiosity ? Or what is the actual use case behind this ?

stray frost Jan 5, 2025, 9:42 PM

#

n some case rouding from negative number give zerro doesnt have the same meaning from rounding a positive number to zero.

#

on you thermometer, you see some time -0 so it might means the real value is -0.001

#

not 0 or +0.001

#

so it might freeze or not ...

grizzled drift Jan 5, 2025, 9:43 PM

#

In that case, is there anything preventing you from checking the number before it gets rounded then ?

stray frost Jan 5, 2025, 9:44 PM

#

yes i need to rewrite alot of stuff for that

grizzled drift Jan 5, 2025, 9:46 PM

#

Erm

stray frost Jan 5, 2025, 9:46 PM

#

I think it is a python behavior not a float behavior ...

grizzled drift Jan 5, 2025, 9:46 PM

#

It's a float behaviour in general, it's defined by a standard

#

https://en.wikipedia.org/wiki/Signed_zero?wprov=sfla1

Signed zero

Signed zero is zero with an associated sign. In ordinary arithmetic, the number 0 does not have a sign, so that −0, +0 and 0 are equivalent. However, in computing, some number representations allow for the existence of two zeros, often denoted by −0 (negative zero) and +0 (positive zero), regarded as equal by the numerical comparison operations ...

stray frost Jan 5, 2025, 9:47 PM

#

they define -0.0 and +0.0 ?

grizzled drift Jan 5, 2025, 9:47 PM

#

Yup, you can read more on that wikipedia page

deft granite Jan 5, 2025, 9:47 PM

#

Yes, IEEE 754 has a few quirks.

stray frost Jan 5, 2025, 9:47 PM

#

but there is no way to test it

grizzled drift Jan 5, 2025, 9:47 PM

#

As for your project I'd have advised for a huge rewrite

deft granite Jan 5, 2025, 9:47 PM

#

str(number).startswith("-") :P

grizzled drift Jan 5, 2025, 9:48 PM

#

But if you relly need to finish it, look into math.copysign

grizzled drift Jan 5, 2025, 9:48 PM

#

deft granite `str(number).startswith("-")` :P

hacky but probably simpler haha

radiant robin Jan 5, 2025, 9:48 PM

#

Just dug through the math module docs. It doesn't seem to offer a way to detect -0.0 vs 0.0. Though 0.0 == -0.0 and 0.0 is not -0.0

#

This relies on 0.0 and -0.0 being interned though.

#

Assuming that, try:

def isnegzero(n):
    return n == 0.0 and n is not 0.0

deft granite Jan 5, 2025, 9:49 PM

#

radiant robin This relies on `0.0` and `-0.0` being interned though.

They're not singletons though

stray frost Jan 5, 2025, 9:49 PM

#

in my present case I will hack to string ....

radiant robin Jan 5, 2025, 9:49 PM

#

Or even:

def isnegzero(n):
    return n is -0.0

grizzled drift Jan 5, 2025, 9:49 PM

#

isn't copysign able to detect negative zeros ?

deft granite Jan 5, 2025, 9:49 PM

#

You can't reliably use is with floats.

radiant robin Jan 5, 2025, 9:49 PM

#

deft granite They're not singletons though

Can you demonstrate that? I'll try.

stray frost Jan 5, 2025, 9:50 PM

#

because I can get a bit more info by using the "sign

steady jolt Jan 5, 2025, 9:50 PM

#

I'd do x==0 and math.copysign(1, x) == -1

deft granite Jan 5, 2025, 9:50 PM

#

radiant robin Can you demonstrate that? I'll try.

!e

x = 0.
y = 3 / float("inf")
print(f"{x is y=}")

mental hawkBOT Jan 5, 2025, 9:50 PM

#

deft granite !e ```py x = 0. y = 3 / float("inf") print(f"{x is y=}") ```

:white_check_mark: Your 3.12 eval job has completed with return code 0.

x is y=False

steady jolt Jan 5, 2025, 9:51 PM

#

On platforms that support signed zeros, copysign(1.0, -0.0) returns -1.0.
Notes like these always scare and confuse me. What kind of platform can you run python on where signed zeros don't exist??

radiant robin Jan 5, 2025, 9:51 PM

#

deft granite !e ```py x = 0. y = 3 / float("inf") print(f"{x is y=}") ```

Thank you.

radiant robin Jan 5, 2025, 9:51 PM

#

steady jolt > On platforms that support signed zeros, copysign(1.0, -0.0) returns -1.0. Note...

ints? 🙂

deft granite Jan 5, 2025, 9:52 PM

#

steady jolt > On platforms that support signed zeros, copysign(1.0, -0.0) returns -1.0. Note...

None that CPython runs on, but Python-the-language doesn't require IEEE 754.

stray frost Jan 5, 2025, 9:53 PM

#

print(math.copysign(1, -0.0))

#

gives -1

grizzled drift Jan 5, 2025, 9:54 PM

#

Sounds good, now you can check with a standard inequality

stray frost Jan 5, 2025, 9:55 PM

#

interesting kirky thing ....

#

... yeah it might not be for all implementation of python

#

thanks all 🙂

grizzled drift Jan 5, 2025, 10:01 PM

#

0.0 is the same as +0.0 for copysign, so you should only get 1(.0 ?) or -1(.0 ?) from it

stray frost Jan 5, 2025, 10:02 PM

#

ya I will used copysign ....

grizzled drift Jan 5, 2025, 10:03 PM

#

I was wondering if you could somehow get "no sign" (i.e. 0 as an output) in this case, but it makes sense that it's not possible here

stray frost Jan 5, 2025, 10:04 PM

#

there is a pep about this

#

https://discuss.python.org/t/accepting-pep-682-format-specifier-for-signed-zero/14088

Discussions on Python.org

Accepting PEP 682 (Format Specifier for Signed Zero)

As the designated PEP-Delegate for John Belmonte’s PEP 682, I’m hereby accepting PEP 682 for inclusion into Python, targeting Python 3.11. I’d like to thank John for a well-written PEP, and for his patience with the PEP process. I’d also like to thank the PEP editors for their invaluable help with the process. I look forward to a slightly more h...

#

is it release i didnt try ...

#

thanks all

mental hawkBOT Jan 5, 2025, 10:08 PM

#

Python help channel closed

This help channel has been closed and it's no longer possible to send messages here. If your question wasn't answered, feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.

#

🔒 zero rounding of float

mental hawkBOT Jan 5, 2025, 11:08 PM

#

Python help channel closed

This help channel has been closed and it's no longer possible to send messages here. If your question wasn't answered, feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.

#🔒 🔒 zero rounding of float