フィルターのクリア

A bug? input: 1/0 output: -Inf

5 ビュー (過去 30 日間)
qilin
qilin 2024 年 2 月 13 日
コメント済み: Matt J 2024 年 2 月 13 日
if x=0, expected output of 1/x is Inf, but matlab output -Inf
load('x.mat')
x==0
Unrecognized function or variable 'x'.
1/0
1/x
  2 件のコメント
Matt J
Matt J 2024 年 2 月 13 日
Your attached file doesn't contain a variable called x.
Image Analyst
Image Analyst 2024 年 2 月 13 日
x==0 is a comparison, not an assignment. Since you didn't assign x yet, you can't see if it's equal to zero, hence the error. I think you meant x=0, not x==0.

サインインしてコメントする。

採用された回答

Stephen23
Stephen23 2024 年 2 月 13 日
編集済み: Stephen23 2024 年 2 月 13 日
"A bug? input: 1/0 output: -Inf"
Not a bug. Because your value x is actually negative zero not positive zero:
Lets try it now:
x = -0
x = 0
1/x
ans = -Inf
num2hex(x) % yep, definitely negative zero
ans = '8000000000000000'
x = +0
x = 0
1/x
ans = Inf
num2hex(x) % yep, that is positive zero
ans = '0000000000000000'
So far everything is working exactly as expected. Note that a few simple arithmetic operations will convert negative zero to positive zero without affecting any other values, e.g. adding zero:
x = -0;
1/x
ans = -Inf
x = 0+x;
1/x
ans = Inf
  4 件のコメント
Dyuman Joshi
Dyuman Joshi 2024 年 2 月 13 日
How did you obtain that value?
Stephen23
Stephen23 2024 年 2 月 13 日
"maybe the next step is to figure out why my ''deltaX'' is negative zero"
Most likely it makes zero(!) difference: note that by definition negative and positive zeros have the same value, so if you are happy with your array being zero-value then simply multiply your array by one (or add zero) and move on to more important tasks.

サインインしてコメントする。

その他の回答 (1 件)

Matt J
Matt J 2024 年 2 月 13 日
編集済み: Matt J 2024 年 2 月 13 日
Theorem: 1/0 = Inf if and only if 1/0 is also -Inf.
Proof: Assume first that 1/0=Inf and let us deduce the implication that 1/0=-Inf. By multiplying the numerator and denominator by -1, we obtain,
1/0 = (-1)/(-0) = (-1)/0 = -(1/0) = -Inf
proving the implication Now assume that 1/0=-Inf. Proceeding similarly,
1/0 = (-1)/(-0) = (-1)/0 = -(1/0) = -(-Inf) = +Inf
and the reverse implication is also proved. Q.E.D.
  2 件のコメント
qilin
qilin 2024 年 2 月 13 日
Thank you for your answer, it's interesting, but: .
In other words, when we say , actually it's , here , so the second
''='' of ''1/0 = (-1)/(-0) =...'' should be modified.
Matt J
Matt J 2024 年 2 月 13 日
You took it way too seriously...

サインインしてコメントする。

カテゴリ

Help Center および File ExchangeFunctions についてさらに検索

タグ

製品


リリース

R2023b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by