CODEWITHSUNDEEP
c#
Trong Tran
Trong Tran

Reputation: 686

Wrong when I convert ulong -> double -> ulong with large number

If small number then it worked, but with larger number then it was wrong. My example code below.

    public void TestZZZZZZZZZ() 
    {
        ulong val = ulong.MaxValue;  // val = 18446744073709551615
        string s = string.Format("{0}", val);

        double d = Convert.ToDouble(s);
        ulong result = ((ulong)d;  // result = 0 <-- WRONG

        Assert.AreEqual(val, result);
    }

Some test data result:

Is there any my wrong? please help.

Thanks!

Upvotes: 0

Views: 1847

Answers (4)

Trong Tran
Trong Tran

Reputation: 686

My solution here was use decimal to contain value for converting (instead double).

Thanks for all your supporting! Trong.

Upvotes: 0

user1968030
user1968030

Reputation:

From Msdn:

Long:Signed 64-bit integer

Double:64-bit floating (15-16 digits Precision)

So result is not same. So if you have just 16 digits your assert is ok. Like M.kazem Akhgary answer.

Upvotes: 2

M.kazem Akhgary
M.kazem Akhgary

Reputation: 19149

double Have precision of 16 digits. So test your number for 16 digits.

ulong val = 1234567890123456; // 16 digits
double d = val;
ulong result = (ulong)d;

Console.WriteLine(val == result); // prints true

After this you will loose precision.

ulong val = 12345678901234567; // 17 digits
double d = val;
ulong result = (ulong)d;

Console.WriteLine(val == result); // prints false

Upvotes: 2

Sergej Christoforov
Sergej Christoforov

Reputation: 1606

That's because long holds 64 bits of data, and significant precision of double is only 53 bits.

Upvotes: 7

Related Questions