Why is 0 divided by 0 throwing an overflow error i

2020-02-06 05:42发布

问题:

Why is 0/0 throwing Overflow error in VBA, while in .Net languages it is simply a Division by 0 error?


E.g., in C# it is a System.DivideByZeroException

static void Main()
{
    int k = 0;
    int p = 0;
    Console.WriteLine(k/p);
}

Div/0 error exists in VBA. But 0/0 gives an overflow exception, while anything else divided by 0 gives a Div/0 exception:

Public Sub TestMe()

    'Integer
    PrintAndCheck (11)      '- Division by zero error

    'Double
    PrintAndCheck (0.9)     '- Division by zero error

    'Long
    PrintAndCheck (50000)   '- Division by zero error

    'String
    PrintAndCheck ("1.1")   '- Division by zero error

    '----------------------------------------------------
    '----------------BUT---------------------------------
    '----------------------------------------------------

    'Integer
    PrintAndCheck (0)       '- Overflow?

End Sub

Public Sub PrintAndCheck(lngDivisor As Variant)

    On Error Resume Next

    Debug.Print lngDivisor / 0
    Debug.Print Err.Description & " from type -> " & VarType(lngDivisor)

    On Error GoTo 0

End Sub

That's what you get in the immediate window:

Division by zero from type -> 2
Division by zero from type -> 5
Division by zero from type -> 3
Division by zero from type -> 8
Overflow from type -> 2

Edit: To make the whole story more interesting:

Public Sub TestMe()
    On Error Resume Next
    Debug.Print Evaluate("0/0")     'Division by 0 error (CVErr(xlErrDiv0)=2007)
    Debug.Print 0 \ 0               'Division by 0 error
    Debug.Print Err.Description
    On Error GoTo 0
End Sub

回答1:

https://docs.microsoft.com/en-us/dotnet/visual-basic/language-reference/operators/floating-point-division-operator

Besides the obvious differences in implementation of languages and the way VBA handles division, MS Doc link above expands on the reasons for overflow exception , that if the operand data types are integer then it will throw Overflow exception (Last statement below)

Alternatively there is \ division operator that checks the range for you and throws Division by zero exception



回答2:

Will try to summarize the answers from the comments:

  1. In VBA 0/0 throws an Overflow exception, because 0/0 is a specific case of division by 0. Thus, it is a good idea to throw a different exception than the standard Division by zero error.

  2. In VBA Evaluate("0/0") returns a Division by zero error, because Evaluate does not raise an error, it returns a Variant value with an error flag, and there is no "overflow" flag available.

  3. In VBA integer division 0\0 returns a Division by zero error, because the result should be an integer value and #IND is a floating point value. As far as #IND cannot be returned, it gives the next best thing - Division by zero error.

More reading concerning 0/0 in other languages:

  • How to produce a NaN float in c?

  • VB.NET